docs(records): row 12 live round and close, queue-commit message follow-up (#1508)

Co-Authored-By: Claude Opus 5.5 <[email protected]>

BUILD-LOG.md

@@ -3365,3 +3365,25 @@ export the suites passed: config 24, task 90, foundation 44, conductor
 first export run failed task 31 and release 3 because Docker had no
 address pools left. Sage removed 12 empty leftover export networks, and
 the rerun passed.
+
+## 2026-09-27: Row 12 done, Piece D's live review round (#1508)
+
+Row 12's gate asked for one review round that writes no new file under
+reviews/, and this round met it.
+
+- Darkwing moved the row to in-review with `--candidate f539466f`, using
+  their own token file. The request posted on the first try as comment
+  26577 on #1508 (rev 15 and 16, commit 3c9f4244).
+- Filbert checked the candidate with `review verify-commit` and posted an
+  approval as comment 26579 with Filbert's own token. `review record`
+  logged it at rev 17.
+- Sage moved the row to done at rev 19 (ff61532a).
+- Nothing was added under docs/plans/reviews/.
+- The queue tests passed 142/142 after the round. The cold-start flake
+  didn't recur.
+
+Correction: the rev 17 entry landed inside f3f48cfd, Darkwing's row 13
+commit. That commit's queue-commit ran between Filbert's record and
+Filbert's own queue-commit, and its message doesn't mention rev 17.
+Filbert's queue-commit then refused with a misleading message, now on
+DEFERRED.

docs/SESSIONS.md

@@ -439,3 +439,5 @@ are never rewritten or removed; corrections are new entries.
 2026-09-27T02:43:14Z | Filbert (T3 Claude Code, thread 9cb9731e) | queue plan C1 correction (#1508, row 12) | per Sage's ruling: plan table line 1470 and 8.9 receipts now gate in-review→waiting-on-jason on reviewer approvals; log entry added; plan sha256 293747cd; uncommitted, for the D commit
 2026-09-27T03:12:01Z | Filbert (T3 Claude Code, thread 9cb9731e) | queue Piece D review r2 (#1508, row 12) | approved: C1 fix correct; build.patch 8a19f7fa manifest 10/10 at cdcedb27, 142/142 tests; 7 of 8 own mutants killed, the stale-approval survivor gets a checked 4-line test (n1); report cf1d3fd0; no commit
 2026-09-27T15:07:51Z | Sage (T3 Claude Code, thread 1ef1e4f8) | queue Piece D commit (#1508, row 12) | manifest b402fb38 19/19; helper lead check (decision 39); nine suites green on export after removing 12 empty leftover networks; f539466f; TOOLS.md from Darkwing's proposal; DEFERRED n2 and cold-start flake
+2026-09-27T15:13:21Z | Filbert (T3 Claude Code, thread 9cb9731e) | row 12 live review round (#1508) | approve posted as comment 26579 with filbert's own token (GET user filbert), recorded rev 17 (filbert-12-record-1); f539466f passes verify-commit; rev 17 committed inside f3f48cfd by a concurrent queue commit; my queue-commit.sh refused at the step7 canary because HEAD moved (misleading message, follow-up); nothing under docs/plans/reviews/
+2026-09-27T15:14:35Z | Sage (T3 Claude Code, thread 1ef1e4f8) | row 12 closed (#1508) | pushed 3c9f4244 after verify --current; Filbert approved via comment 26579 (rev 17); row 12 done rev 19 ff61532a; DEFERRED queue-commit HEAD-moved message

docs/plans/DEFERRED.md

@@ -151,6 +151,16 @@ at every gate. Started 2026-09-12 during the control board MVP.
   That isn't confirmed. Next time: capture the name, then fix it as its
   own patch. (2026-09-27, #1508)
 
+- **queue-commit.sh blames the guard when HEAD moves.** Step 7's canary
+  reads H into its index (`scripts/queue-commit.sh:94`). The hook diffs
+  against the live HEAD. If another queue commit moves the branch after H
+  is recorded, the clean canary sees a difference and refuses with "git is
+  not running the queue guard as installed". The fix it prints does
+  nothing. The refusal is correct, the message is wrong. Fix: guard_check
+  compares HEAD with H before the canary and says "HEAD moved since H;
+  start again". Filbert found this during the row 12 live round.
+  (2026-09-27, #1508)
+
 ## Queue
 
 Moved to `docs/plans/QUEUE.md` on 2026-09-13. This file holds only gaps.