Block a CAT-UNVERIFIED tag that hands a read-only check back as 'say X'

[EdbertChan]

Summary

Small checker scripts read each reply before Claude sends it. One checks that claims come with proof.

The problem: Claude could mark a claim "unverified" and then ask the user to say "test it" for a check Claude can run itself.

The fix: that reply is now refused and Claude is told to run the check. A second checker now accepts "I am running it now" as a valid next step.

Offers tied to real actions, like pushing or deleting, still pass.

Review Claim

A reply that hands a harmless check back to the user is refused, and a reply saying the check is running now is accepted.

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 is the first slice: it adds a new refusal rule, and the later slices reuse the same ledger checker. It is one claim: stop handing checks back. The other slices change wording, retry handling, and citations.

Non-goals

• Does not change when a tag is accepted after a real attempt.
• Does not touch offers that name a write such as push, deploy, or delete.
• Does not change the word-count or claim checks.

Measured

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

• base: frustration-watchdog: Ran 25 tests, FAILED (failures=1); unverified-tag-ledger: Ran 57 tests, FAILED (failures=3)
• head: frustration-watchdog: Ran 25 tests, OK; unverified-tag-ledger: Ran 57 tests, OK

Test Plan

Test Plan

• source scripts/test/hermetic_env.sh; python3 engine/hooks/frustration-watchdog/tests/test_hooks.py -> Ran 25 tests, OK
• source scripts/test/hermetic_env.sh; python3 engine/hooks/unverified-tag-ledger/tests/test_hooks.py -> Ran 57 tests, OK
• python3 scripts/ci/check_hook_test_coverage.py engine/hooks/frustration-watchdog -> check_hook_test_coverage: OK (1 hook(s) checked)
• python3 scripts/ci/check_hook_test_coverage.py engine/hooks/unverified-tag-ledger -> check_hook_test_coverage: OK (1 hook(s) checked)
• Positive and negative cases are in the new tests: a "say test it" offer fires, an offer tied to a write stays silent.

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

[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: 47dcc35c-5560-404c-b2c0-6099c0cb0852)

[mergify]

Queued — the merge queue status continues in this comment ↓.

[EdbertChan]

@Mergifyio queue

[mergify]

Merge Queue Status

• ✅ Entered queue — 2026-10-13 10:13 UTC · Rule: admin-bypass · triggered by a @mergifyio queue command
• 🟠 Checks running · on draft merge queue: checking #1378 + #1381 together on main (f497aa7) #1397
• 🚫 Left the queue — 2026-10-13 10:18 UTC · at 1901980b22a5b0555dcb31c9c8a8a6a41fb5048f

This pull request spent 4 minutes 22 seconds in the queue, with no time running CI.

Reason

Pull request #1381 has been merged manually at 26a533a

Hint

You were too fast!

Waiting for

• check-success = test

All conditions

• check-success = test
• check-success = lint
• check-success = validate

Tick the box to put this pull request back in the merge queue (same as @mergifyio queue).

• Requeue this pull request

[mergify]

queue

☑️ The pull request has been merged manually

Details

The pull request has been merged manually at 26a533a