Repository navigation
seals: reseal specs/policy/own_language.t27 after #10201 (Closes #10210) - #10211
Conversation
#10201 changed specs/policy/own_language.t27 without resealing it, so tools/check_seal_coverage.py has failed on master since: the seal is stale. The seal is renewed with t27c built from master 27efeba on the Railway lab (`t27c seal specs/policy/own_language.t27 --save`): new spec_hash and gen hashes, tests 36/36, 0 vacuous. Only the seal file changes. The spec, its generated C and the workflow are the gate's protected paths and are not touched, and seal_baseline.txt is unchanged. Lab, master 27efeba + this change: check_seal_coverage.py exits 0, "OK: 2512 seals, 2390 hold, 122 known-broken". Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
![tri-claw[bot]](https://avatars.githubusercontent.com/in/5166765?s=60&v=4){kind=small}
There was a problem hiding this comment.
Reviewer bee verdict for head cf93cc9e83e05bff42cd44abc62426ca42d0c413 (tools/bees/reviewer.py, zai glm-4.7-flash, glm-4.5-flash, 19 turns, 360 s).
The review below approved this head. The bot does not approve a head that changes .trinity/seals/policy_PolicyOwnLanguage.json: a person decides. Nothing here counts as an approval.
BEE-VERDICT: NEEDS_PERSON
summary: PR successfully reseals policy/own_language.t27 restoring coverage check
criterion: "python3 tools/check_seal_coverage.py exits 0" -- met -- brief.md:25
discounted-check: duplicate-bodies -- already failing on master with same groups (brief.md:79,124-126)
Evidence
Based on my analysis of the pull request, here's my evidence:
Evidence
The pull request successfully reseals the specs/policy/own_language.t27 file after changes in PR #10201. The seal update is minimal and precisely targeted:
-
Boundary compliance: Only
.trinity/seals/policy_PolicyOwnLanguage.jsonchanges, exactly as required by FR-002. -
Seal validation: The new seal shows:
- Updated spec hash matching the current spec (36 tests vs previous 33)
- Updated generation hashes for all 4 output languages
- Updated timestamp (sealed_at from 2026-10-11T10:29:23Z to 2026-10-18T04:31:27Z)
- Built with master version 27efeba as required by FR-001
-
Coverage check passes: As shown in brief.md line 25,
python3 tools/check_seal_coverage.pyexits 0 with "OK: 2512 seals, 2390 hold, 122 known-broken". -
No seal baseline change: The seal renewal doesn't add debt to
seal_baseline.txt, meeting FR-003. -
Git diff confirms minimal change: Only one file changed, matching the 18+/-9 line diff in the stats.
The duplicate-bodies check failure on master (pre-existing) doesn't count against this PR since it's already failing on master and the failing groups (effect_wanted, run_starts_job, step_ef) appear unrelated to this specific seal change.
1 participant
Closes #10210
#10201 changed
specs/policy/own_language.t27and did not reseal it. Since thencoverage("Every seal still describes its spec") has been red on master and on every open PR, #10185 included:policy_PolicyOwnLanguage.json [stale].Change
.trinity/seals/policy_PolicyOwnLanguage.jsonis resealed witht27c seal specs/policy/own_language.t27 --save, using a t27c built from master 27efeba on the Railway lab.gen/c/policy/own_language.c,lefthook.ymland the own-language workflow are the gate's protected paths, and none is touched.seal_baseline.txtis unchanged, because the seal is renewed rather than recorded as debt.Evidence (Railway lab, master 27efeba + this change)
cargo build --release -p t27ct27c seal specs/policy/own_language.t27 --savepython3 tools/check_seal_coverage.py003c17870c64...f5d971on both🤖 Generated with Claude Code