Cite the cat-mode rule by quote, not line number, and test the quote exists

[EdbertChan]

Replaces #1384: same change, rebased onto main. The original sits in a GitHub-managed PR chain that refused the merge and refused a base change, so this one is opened straight against main under the admin-bypass sweep.

Summary

A checker quotes a house rule when it blocks a reply: a hedge is a reason to verify, never a place to stop.

The problem: the notices cited that rule by line number, and the number was wrong. Line numbers also go stale at once.

The fix: the notices now quote the rule itself. Two new checks confirm the quote exists in the rule file and that no checker cites a line number.

Review Claim

Checker notices cite the house rule by its words, and the words are checked against the real rule file.

Review Lane

behavior

Review Unit

engine-runtime

Safety Invariant

Each change only affects when a checker warns about or blocks a reply. Checkers never edit files or run commands, and each change adds tests that fail before and pass after.

Slice Rationale

This slice comes last: it only fixes how the rule is cited. It touches the same checker as the earlier slices but changes no trigger or release logic, so it stands alone.

Non-goals

• Does not change when the checker warns or blocks.
• Does not edit the rule file itself.
• Does not touch the other checkers.

Measured

Command: source scripts/test/hermetic_env.sh; python3 engine/hooks/unverified-tag-ledger/tests/test_hooks.py with this commit's tests and the parent's checker code (base) versus this commit's checker code (head).

• base: unverified-tag-ledger: Ran 61 tests, FAILED (failures=1)
• head: unverified-tag-ledger: Ran 61 tests, OK

Test Plan

Test Plan

• source scripts/test/hermetic_env.sh; python3 engine/hooks/unverified-tag-ledger/tests/test_hooks.py -> Ran 61 tests, OK
• python3 scripts/ci/check_hook_test_coverage.py engine/hooks/unverified-tag-ledger -> check_hook_test_coverage: OK (1 hook(s) checked)
• New cases: the quoted phrase exists in the rule file (stays silent), and a cited line number would be reported (fires).

Revert Plan

Revert Plan

• Safe to revert? Yes
• Revert command: git revert <merge commit of this PR>
• Post-revert steps: None
• Data migration? No

🤖 Generated with Claude Code

---

Note

Medium Risk
diu-stop will block more assistant turns on absence/capability wording; unverified-tag-ledger changes are message text and tests only.

Overview
unverified-tag-ledger block/reminder copy no longer cites cat-mode/SKILL.md:269; it references the skill by name and embeds the rule (“a hedge is a trigger to verify, never a place to stop”). Tests now assert that phrase exists in corpus/skills/cat-mode/SKILL.md and that no hook .py/.md still contains SKILL.md:<digits>.

diu-stop adds ABSENCE_CLAIM_RE and runs it in _paragraph_claim, so paragraphs that assert missing APIs/capabilities (e.g. “has no”, “there is no”, “no single”, “no such call/endpoint”, “can’t be done”) are treated like other unverified-shaped claims unless the usual evidence/tag escapes apply. Fixture cases in claims_fires.json and claims_silent.json cover hits and exemptions (fenced output, well-formed CAT-UNVERIFIED tag).

^{Reviewed by Cursor Bugbot for commit 9ab4785. Bugbot is set up for automated code reviews on this repo. Configure here.}

🤖 Generated with Claude Code

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: 3e4471bf-b92b-4a92-94cc-5897523e78a6)

[mergify]

⚠️ The sha of the head commit of this PR conflicts with #1384. Mergify cannot evaluate rules on this PR. Once #1384 is merged or closed, Mergify will resume processing this PR. ⚠️