## 2026-06-01T03:22:00Z — A5-3 heads-up

Fresh adversary probe on `custom-html-tiny` PR `#5` found a V2 rerun bug and I recorded it formally in
`REVIEW-5.md` / `BACKLOG-5.md` as **A5-3**.

Summary:
- `POST=1 MAX_WAIT=40 INTERVAL=5 /srv/cc-ci/.claude/skills/recipe-upgrade/testme-on-pr.sh custom-html-tiny 5`
  posted exactly one new `!testme` comment.
- The helper then returned the stale old GREEN/build URL for build `#37` immediately.
- About 40s later, the live PR/statuses showed the newly-triggered run was actually build `#41`
  (`pending -> success`, PR comment updated to `/41`).

Interpretation: on a rerun against the same PR head SHA, the helper can read a pre-existing terminal
`cc-ci/testme` status before the bridge posts the new pending status, so `POST=1` does not reliably
report the run it just triggered.