[trissim]

Tonight #3200, #3187 and #3151 each merged an approved old head while its fix was still being written.

The writer already withdraws; nobody called it.

• Every head verdict posts agent-comms/review through the one writer, HeadReview.post (src/agent_comms/head_review.py:54), called from HeadVerdict.reviewed (src/agent_comms/todos.py:3337).
• judge back posts RejectedHeadVerdict.state = "failure", and judge reviewing posts UnderReviewHeadVerdict.state = "pending" (pending since 7dac490, 10-24 06:51).
• GitHub keeps the newest status per context on a commit, so either verb on an ok'd sha withdraws the success.
• The three bad merges happened because the rework started without anyone running judge reviewing on the old head. With no withdrawal posted, auto-merge took the old head.

The test (ok then back on one sha was already covered at tests/test_head_review.py:59):

• ok then reviewing on the same sha posts state=success, then state=pending.
• Changing UnderReviewHeadVerdict.state to success makes it fail.
• tests/test_head_review.py: 8 passed.

🤖 Generated with Claude Code

https://claude.ai/code/session_013ejYrMLnLWah8wCDDcb57V

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota.

An organization owner can choose the repository organization as the billing source for reviews requested by members with a paid Copilot license. Learn about Copilot code review billing.