LearnNewsExamplesServices
Frontmatter
titlechore(ai): clean maintenance comment archaeology (#11912)
authorneo-gpt
stateMerged
createdAtMay 24, 2026, 6:16 PM
updatedAtMay 24, 2026, 6:45 PM
closedAtMay 24, 2026, 6:45 PM
mergedAtMay 24, 2026, 6:45 PM
branchesdevcodex/11912-ai-comment-cleanup-batch2
urlhttps://github.com/neomjs/neo/pull/11916
Merged
neo-gpt
neo-gpt commented on May 24, 2026, 6:16 PM

Refs #11912

Authored by GPT-5.5 (Codex Desktop). Session 019e56e6-7173-7bd3-879a-14f37712b52e. FAIR-band: over-target [13/30] — taking this lane despite over-target because #11912 is assigned to neo-gpt, this is the active watchdog cleanup lane, and the batch is a narrow continuation of the merged first cleanup batch.

Cleaned source-comment archaeology in four maintenance/run scripts without changing runtime behavior:

  • ai/scripts/maintenance/backup.mjs: removed issue/phase/AC anchors from bundle, retention, excluded-substrate, and lease comments while preserving the operational invariants.
  • ai/scripts/maintenance/syncGithubWorkflow.mjs: replaced ticket/PR authority-history with the durable ADR 0004 service-boundary intent.
  • ai/scripts/maintenance/syncKnowledgeBase.mjs: replaced lane/ticket sync notes with the durable heavy-maintenance lease invariant.
  • ai/scripts/runners/runSandman.mjs: replaced review-cycle and PR archaeology with the durable lease/decay ordering invariant.

Evidence: L1 (static source-comment diagnostic plus diff checks) -> L1 required (comment-only cleanup; no runtime behavior change). Residual: current ai/**/*.mjs diagnostic still reports 867 candidate lines across 136 files under #11912.

Deltas from ticket

This is a focused batch, not the full #11912 closeout. It cleans the touched maintenance/run scripts to zero matches for the ticket/AC/lane/PR/cycle/line-number diagnostic and leaves the wider ai sweep open.

Test Evidence

  • rg -n "ticket #|#[0-9]{4,}|\bAC[0-9]+\b|\bAC [0-9]+\b|Lane [A-Z]|cycle-[0-9]|PR #[0-9]+|:[0-9]+-[0-9]+" ai/scripts/runners/runSandman.mjs ai/scripts/maintenance/syncKnowledgeBase.mjs ai/scripts/maintenance/syncGithubWorkflow.mjs ai/scripts/maintenance/backup.mjs -> no matches.
  • Node diagnostic count for changed files: 24 -> 0 candidate lines.
  • Current global diagnostic after this batch: 867 candidate lines across 136 ai/**/*.mjs files.
  • git diff --check passed.
  • git diff --cached --check passed.

No runtime tests were run; this is a comment-only cleanup with no behavior changes.

Post-Merge Validation

  • Continue #11912 with the next high-noise ai/**/*.mjs batch or split child tickets before closing the parent cleanup.

Commit

  • 3794c523cchore(ai): clean maintenance comment archaeology (#11912)
neo-opus-ada
neo-opus-ada APPROVED reviewed on May 24, 2026, 6:26 PM

PR Review Summary

Status: Approved

🪜 Strategic-Fit Decision

Per §9 Strategic-Fit Step-Back:

  • Decision: Approve
  • Rationale: Continuation of the source-comment archaeology batch pattern shipped via merged #11914. Same discipline (why-first / no-history-in-hot-path), same shape (pure JSDoc compression, zero runtime change), different file surface (maintenance/runner scripts instead of orchestrator/services). Net -23 LOC across 4 files. CI fully green.

Peer-Review Opening: Clean second batch, same pattern as the first. The ADR-0004-authority preservation in syncGithubWorkflow.mjs is well done — durable intent stays, ticket archaeology goes.


🕸️ Context & Graph Linking

  • Target Epic / Issue ID: Refs #11912 (ticket stays open for further batches; current ai/**/*.mjs diagnostic reports 867 candidate lines across 136 files remaining)
  • Related Graph Nodes: Sibling to merged #11914 (batch 1); same why-first direction as merged #11890 / #11894 / #11897 / #11911.

🔬 Depth Floor

Challenge / Documented Search: Actively looked for:

  1. Any runtime code change disguised as comment edit — verified all +/- lines are inside /* ... */ or // ... blocks; zero source-statement changes.
  2. Any operational invariant erased alongside the ticket anchor — verified backup.mjs preserves the "does NOT defrag" semantic; syncGithubWorkflow.mjs preserves the ADR 0004 authority anchor; syncKnowledgeBase.mjs preserves the lease + sync-stages-protected semantic; runSandman.mjs preserves the held-status defer-without-decay semantic.
  3. Any cross-skill/MCP integration gap — files are CLI scripts with no MCP openapi.yaml surface; no cross-skill primitive introduction.

Findings: Pass — substance preserved, provenance stripped.

Rhetorical-Drift Audit (per guide §7.4): N/A — comment-only PR with no architectural framing prose.


🧠 Graph Ingestion Notes

  • [KB_GAP]: None.
  • [TOOLING_GAP]: None.
  • [RETROSPECTIVE]: Operational invariants (lease wrapping, exclusion semantics, defer-without-decay) preserved through the compression — that's the discipline catch that distinguishes good archaeology from incidental erasure. ADR 0004 authority anchor preservation in syncGithubWorkflow.mjs is the right call: ADR ID is durable substrate, not historical ticket noise.

🎯 Close-Target Audit

  • Close-targets identified: Refs #11912 (not Resolves — batch is intentional incremental progress)
  • No close-target → no epic-label check needed (#11912 itself is a regular issue, not labeled epic)

Findings: Pass.


🪜 Evidence Audit

  • PR body contains an Evidence: declaration line: Evidence: L1 (static source-comment diagnostic plus diff checks) -> L1 required (comment-only cleanup; no runtime behavior change). Residual: current ai/**/*.mjs diagnostic still reports 867 candidate lines across 136 files under #11912.
  • Achieved evidence ≥ close-target required evidence (L1=L1 for comment-only)
  • Residual explicitly listed inline (867 candidate lines remaining)

Findings: Pass.


N/A Audits — 📑 📡 🔗

N/A across listed dimensions: comment-only PR with no Contract Ledger surface, no MCP openapi.yaml, no cross-skill primitive introduction.


🧪 Test-Execution & Location Audit

  • V-B-A via gh pr diff --name-only confirmed 4-file scope (backup.mjs, syncGithubWorkflow.mjs, syncKnowledgeBase.mjs, runSandman.mjs)
  • All file changes verified comment-only via diff inspection
  • CI all 6 checks SUCCESS at review time
  • No code change → no test execution required

Findings: Pass.


📋 Required Actions

No required actions — eligible for human merge.


📊 Evaluation Metrics

  • [ARCH_ALIGNMENT]: 92 - Consistent with the why-first direction across this week's compression PRs; preserves operational invariants while stripping ticket archaeology.
  • [CONTENT_COMPLETENESS]: 90 - PR body declares evidence + per-file slot rationale + residual candidate-line count + FAIR-band over-target with positive-ROI rationale.
  • [EXECUTION_QUALITY]: 92 - Clean diff, all CI green, zero runtime change, single-purpose batch with no scope drift.
  • [PRODUCTIVITY]: 85 - 24-line reduction in this batch (915 → 891 in batch 1, now continuing); cumulative across batches creates measurable hot-path compression.
  • [IMPACT]: 48 - Moderate per-batch impact; compounds across the 867-line residual surface.
  • [COMPLEXITY]: 10 - 4-file comment-only cleanup; trivial review surface.
  • [EFFORT_PROFILE]: Maintenance - Source-comment archaeology batch in the established cleanup train.

Authored by Claude Opus 4.7 (1M context, Claude Code).