Authored by Claude Fable 5 (Claude Code). Session 567191c3-16f6-4235-914c-b51fc94d1514.

Resolves #12876 Related: #12696

Refreshes the v13 release note's at-cut evidence surfaces, per the Post-Merge Validation checklist of merged PR #12697: live release-window counts (1,260→1,307 merged PRs, 1,664→1,717 closed issues since v12.1.0, verified 2026-06-11), the curated-digest mermaid corpus figures (1,200+/1,600+ → 1,300+/1,700+), and the maintainer contribution table (ada 2,003→2,017, claude-opus 201→223, vega 137→159, fable "newly onboarded"→25; gpt/gemini unchanged — method-reproduction confirmed the original derivation). The appendix snapshot bullets now carry explicit freshness boundaries (live vs lagging mirror) instead of "to be refreshed once at final cut".

Evidence: L1 (static content verification: live GraphQL count probes, per-user contributionsCollection GraphQL, appendix generator re-run → 2,933 lines, post-edit grep sweep for stale values) → L1 required (documentation evidence refresh; no runtime ACs). No residuals — the AC3 operator runbook is posted: https://github.com/neomjs/neo/issues/12696#issuecomment-4675564872

Deltas from ticket

• README / learn/benefits/AIEngineeringTeam.md deliberately NOT touched. Mid-implementation V-B-A surfaced that #12725 (the prior identity-count refresh attempt) was operator-rejected on 2026-06-08: public identity surfaces show fixed-month velocity proof ("In May 2026 … 706/800"), not drifting release-window stats. The release note's appendix snapshot is the opposite class — release-window evidence inside the release artifact itself, with an explicit refresh-at-cut instruction — so only the note changes here.
• Maintainer-table derivation verified before writing: GraphQL contributionsCollection.contributionCalendar.totalContributions reproduces the documented values exactly for the two inactive accounts (gpt 1,510, gemini 1,039), confirming the method; active accounts refreshed with the same query.
• One prose line adjusted for factual accuracy: @neo-fable's "contribution history starts post-v13" → "first contributions landed in the release's final nights" (25 contributions exist pre-cut; the old claim aged out mid-release).

Test Evidence

• gh api '/search/issues?q=repo:neomjs/neo+is:pr+is:merged+merged:%3E=2026-03-27' → 1307
• gh api '/search/issues?q=repo:neomjs/neo+is:issue+is:closed+closed:%3E=2026-03-27' → 1717
• Per-user contributionsCollection GraphQL → ada 2017, gpt 1510, gemini 1039, claude-opus 223, vega 159, fable 25
• node buildScripts/release/analyzeClosedSinceRelease.mjs 2026-03-27 --format markdown --include-items --output /tmp/v13-release-appendix.md → 2,933 lines; mirror totals 1,228 PRs / 1,587 issues (unchanged mirror boundary, documented as such)
• Post-edit sweep: grep -n "1,260\|1,664\|2,003\|newly onboarded\|1,200+\|1,600+" on the note → zero matches
• Pre-commit check-chore-sync + whitespace lint passed on the agent/sync-* branch; branch freshness: merge-base == origin/dev tip; single commit, no stray close-keywords in branch history

Post-Merge Validation

• At publish: publish.mjs Step 6 runFullSync reconciles the lagging mirror counts (documented in the runbook comment; no action before cut)

Review note

§6.1 micro-change exemption claimed: pure documentation, no runtime impact (12 lines changed in resources/content/release-notes/v13.0.0.md). Cross-family review welcome but non-blocking for the cut; operator reviews the note at publish regardless.