Applies the operator's rulings of 2026-11-08 on the decisions raised in #2.
Ruling
Change
Kasm is not required
Removed from the catalog (now 13 services), the OIDC list, secrets and onboarding order
Drop ACME revocation
R-TLS-02 now reads "issued and renewed". docs/08 states that a compromised key is handled by rotation
No reason for automated fault injection
R-TEST-03 amended; new ADR-0026
Homepage should be behind login
R-IDP-02 was already right, so no text change. The gap stays in the register, marked as ruled, until the implementation lands in a separate PR
Please check how I read the fault-injection ruling. I took it to mean that L3 tests are never started by CI or a schedule, and do not gate individual merges, but still exist and are run by you when you choose. R-TEST-05 is unchanged. If you meant to retire fault-injection testing altogether, that is a further change to R-TEST-05 and the tests.
Consequence accepted in ADR-0026: a merge can change a safety-critical behavior without its L3 test re-running, so an L3 proof can go stale.
ADR-0014's Context still names Kasm, because accepted ADR text is not rewritten. The frozen agents/task-board.md is untouched for the same reason.
Still open: the photos/ question, the one item left at the top of docs/16-known-gaps.md.
After merging:ci/SPEC_PIN moves to this merge commit, in impl PR #160. That PR also picks up the R-TLS-02 registry summary and the stale Kasm line in catalog/README.md.
Checked: requirement ID set unchanged; relative links resolve. Not checked: the spec repo has no CI, and this branch has not been run through the implementation's gate. No live state was measured.
Applies the operator's rulings of 2026-11-08 on the decisions raised in #2.
| Ruling | Change |
|---|---|
| Kasm is not required | Removed from the catalog (now 13 services), the OIDC list, secrets and onboarding order |
| Drop ACME revocation | **R-TLS-02** now reads "issued and renewed". `docs/08` states that a compromised key is handled by rotation |
| No reason for automated fault injection | **R-TEST-03** amended; new **ADR-0026** |
| Homepage should be behind login | **R-IDP-02 was already right, so no text change.** The gap stays in the register, marked as ruled, until the implementation lands in a separate PR |
**Please check how I read the fault-injection ruling.** I took it to mean that L3 tests are never started by CI or a schedule, and do not gate individual merges, but still exist and are run by you when you choose. **R-TEST-05 is unchanged.** If you meant to retire fault-injection testing altogether, that is a further change to R-TEST-05 and the tests.
Consequence accepted in ADR-0026: a merge can change a safety-critical behavior without its L3 test re-running, so an L3 proof can go stale.
ADR-0014's Context still names Kasm, because accepted ADR text is not rewritten. The frozen `agents/task-board.md` is untouched for the same reason.
**Still open:** the `photos/` question, the one item left at the top of `docs/16-known-gaps.md`.
**After merging:** `ci/SPEC_PIN` moves to this merge commit, in impl PR #160. That PR also picks up the R-TLS-02 registry summary and the stale Kasm line in `catalog/README.md`.
**Checked:** requirement ID set unchanged; relative links resolve. **Not checked:** the spec repo has no CI, and this branch has not been run through the implementation's gate. No live state was measured.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Retire Kasm from the catalog. Drop ACME revocation from R-TLS-02. Make L3
fault injection operator-run and exclude it from the merge gate (R-TEST-03,
ADR-0026). Update the gap register to match; Homepage authentication is now
ruled must-enforce and stays a gap until the implementation lands.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Applies the operator's rulings of 2026-11-08 on the decisions raised in #2.
docs/08states that a compromised key is handled by rotationPlease check how I read the fault-injection ruling. I took it to mean that L3 tests are never started by CI or a schedule, and do not gate individual merges, but still exist and are run by you when you choose. R-TEST-05 is unchanged. If you meant to retire fault-injection testing altogether, that is a further change to R-TEST-05 and the tests.
Consequence accepted in ADR-0026: a merge can change a safety-critical behavior without its L3 test re-running, so an L3 proof can go stale.
ADR-0014's Context still names Kasm, because accepted ADR text is not rewritten. The frozen
agents/task-board.mdis untouched for the same reason.Still open: the
photos/question, the one item left at the top ofdocs/16-known-gaps.md.After merging:
ci/SPEC_PINmoves to this merge commit, in impl PR #160. That PR also picks up the R-TLS-02 registry summary and the stale Kasm line incatalog/README.md.Checked: requirement ID set unchanged; relative links resolve. Not checked: the spec repo has no CI, and this branch has not been run through the implementation's gate. No live state was measured.
🤖 Generated with Claude Code