docs(readiness): re-measure at 1e234a011, clause 4 not green at the tip; D-12 names the Codex session

[Chris0Jeky]

Docs-only, two files. Coordinator re-measure of the v0.3.0 readiness view at main 1e234a011 (2026-09-05 04:00Z) and one human-action item rewritten with what is now known.

docs/releases/V0_3_0_READINESS.md:

• Clause 2: 44 open (from 48 at the 22:50Z count: six closed on evidence, #2460, #2501, #2524, #2563, #2570, #2572; two seeded and open, #2582, #2588; two seeded and closed the same night, #2583, #2604). Section 5 lists regenerated live: 11 dogfooding / 20 ci / 13 other, 20 gated and 24 un-gated, arithmetic stated.
• Clause 4: stated as not green at the tip when measured. Last completed green run is still 330ccb4de (23:19Z). Every later push run on main was cancelled under the next merge; the 1e234a011 run had 13 green jobs and one real failure, the #2378 Windows launcher-suite timeout (fourth distinct case that night, recorded on the issue), before it too was cancelled. Names the two changes that make the clause read green with confidence: #2610 (parked under SC-10) and #2378.
• Clause-4 noise list: #2572 out (closed on PR #2592), #2588 in (Linux dev-up TERM case); #2378's five further occurrences that night listed.
• SC-10 row: ten parked control-plane PRs, adding the beta lane's #2587, #2608 and #2610, each with a SHIP review.

OUTSTANDING_TASKS.md, item D-12 rewritten: the unclaimed PR author is a Codex session posting [Codex lane claim v2] under the alpha lane's name (first claim 03:35Z on #2128); fourteen PRs listed with their issues, two closed as superseded, the interim rule in force (first recorded claim or open PR wins regardless of marker form; the Claude lanes do not adopt or review its PRs; nothing of its merges without the gate's review), and the two answers wanted from the maintainer. No item ticked; nothing inferred.

Verified: node scripts/check-docs-governance.mjs passed; git diff --check clean; milestone lists pulled with gh issue list --milestone at write time; run states read from the Actions API; branch authorship read from git log on the fetched refs.
NOT verified: nothing executable changed.

Refs #2235, #2378, #2582

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[Chris0Jeky]

Review gate (Codex credits exhausted, SC-9): one fresh-context read-only reviewer on head b7a765853 (base main 9cf28bbc6). Verdict: SHIP.

Checked by reading: section 5's lists are disjoint and count 11 / 20 / 13 = 44, the ci sub-lists sum to 20, gated 20 and un-gated 24 reconcile, and the 48 minus 6 plus 2 delta is stated correctly; the closed and seeded sets are identical at the clause-2 row and in the movement paragraph; clause 4, the clause-4 risks paragraph and the section 5 ci list agree on the three open reds (#2378, #2561, #2588) and on #2572 being closed; header, intro and section 5 agree on 1e234a011 at 04:00Z; SC-10 lists exactly ten PRs, all parked, matching OUTSTANDING_TASKS.md; D-12 records no ruling and authorizes no deletion or merge; no em dashes in the new text.

Five LOW findings, none blocking; disposition under law 2 (no fix-commit cascade for LOWs):

• Interim rule in D-12 ("first recorded claim or open PR wins") is broader than the memory file's "open PR plus a current claim outranks a new claim". Declined as written: the broader rule is a coordinator-set interim restraint that only adds protection for unclaimed open PRs while D-12 is unanswered; it lives in the human-action item, not the memory, and lapses with the answer.
• Section K's preamble still scopes the section to the audit and promises a proposed default per item; D-12 is a coordinator observation with none. Tracked for the next OUTSTANDING pass: one clause in the preamble.
• "Two shapes seeded from the same day's reds take its place" now precedes three issue names, one of them closed and one seeded the next day. Tracked for the next readiness re-measure, which rewrites that paragraph anyway.
• "this 2026-09-04 re-measurement" in the movement paragraph is stale deixis, disambiguated by "the morning count" in the same sentence. Same disposition.
• #2562 is counted under section 2 but section 2's body never names it. Same disposition: the next re-measure adds it to the clause-5 chain text.

Merge after ci-required is green at b7a765853 and the head has aged.

[Chris0Jeky]

Addendum after the review: commit 942b37a adds one dated fact to the clause-2 row and the section 5 movement paragraph, #2639 (Priority I security, PWA activate-time re-sweep never runs; seeded 04:19Z, after the 04:00Z measurement; beta lane holds the fix), with the live count stated as 45 and 25 un-gated. No number in the measured snapshot changed and no other text moved, so the SHIP verdict covers it; merge waits for ci-required at 942b37a plus aging.