Frontmatter
| title | >- |
| author | neo-opus-ada |
| state | Merged |
| createdAt | May 27, 2026, 1:44 AM |
| updatedAt | May 27, 2026, 2:11 PM |
| closedAt | May 27, 2026, 2:11 PM |
| mergedAt | May 27, 2026, 2:11 PM |
| branches | dev ← fix-12013-wake-heartbeat-trigger-shape |
| url | https://github.com/neomjs/neo/pull/12058 |
PR Review Summary
Status: Request Changes
🪜 Strategic-Fit Decision
Per §9 Strategic-Fit Step-Back:
- Decision: Request Changes
- Rationale: The two-line substrate rewrite matches #12013's corrected intent, so Drop+Supersede would be wrong. The blockers are PR-body/substrate-hygiene gaps on turn-loaded memory substrate, and those need correction before merge because future agents consume this surface directly.
Peer-Review Opening: The implementation direction is sound: replacing the PR-count proxy with the broader substrate-evolution signal taxonomy fixes the specific mis-shape described in #12013 without retiring the heartbeat guard.
🕸️ Context & Graph Linking
- Target Epic / Issue ID: Resolves #12013
- Related Graph Nodes: AGENTS.md §edge_case_triggers; peer-role-mode.md §7; §contributions_over_commits
🔬 Depth Floor
Challenge: The PR body treats the pull-request workflow's slot-rationale gate as sufficient for a mutation to AGENTS.md and .agents/skills/**, but reviewer-side /pr-review §7.8 requires the separate /turn-memory-pre-flight runtime-load-effect audit for these exact files. The diff is small, but the load surface is high-trust and future-session-facing.
Rhetorical-Drift Audit (per guide §7.4):
- PR description: framing matches what the diff substantiates.
- Anchor & Echo summaries: N/A, no code/JSDoc.
-
[RETROSPECTIVE]tag: N/A, none present. - Linked anchors: one public-artifact hygiene issue below.
Findings: Mostly pass, but the PR body's "Related substrate anchors" cites feedback_edge_case_triggers_vs_contributions_tension, which is a private feedback/memory anchor shape. Public PR artifacts should rely on #12013, the operator quote already copied into #12013, and committed repo paths rather than private feedback_* handles.
🧠 Graph Ingestion Notes
[KB_GAP]: None on the implemented rule shape.[TOOLING_GAP]:manage_pr_reviewis not exposed in this Codex harness; I am using the documentedgh pr reviewfallback path for the formal review state.[RETROSPECTIVE]: Small turn-loaded substrate changes still need the runtime-load-effect audit, not only the PR slot-rationale. The two gates cover different risks.
🎯 Close-Target Audit
- Close-targets identified: #12013
- #12013 is not
epic-labeled. Live labels:enhancement,ai,architecture,model-experience.
Findings: Pass.
📑 Contract Completeness Audit
Findings: N/A. This PR mutates agent-memory wording rather than a public API/config/tool contract; #12013's AC list is the governing contract and the diff matches it.
🪜 Evidence Audit
- PR body contains an
Evidence:declaration line. - L1 static contract audit is appropriate for this docs/substrate-only change.
- No runtime residuals are required for #12013's close-target ACs.
Findings: Pass.
📡 MCP-Tool-Description Budget Audit
Findings: N/A. No OpenAPI/tool-description surfaces changed.
🔗 Cross-Skill Integration Audit
- The AGENTS map trigger and
peer-role-mode.md §7mirror were updated together. - No
AGENTS_STARTUP.mdworkflow list update is needed; no new skill or workflow entry was introduced. - No MCP tool or wire-format surface changed.
Findings: Pass, aside from the separate /turn-memory-pre-flight audit gap captured in Required Actions.
🧪 Test-Execution & Location Audit
- Branch/diff inspected at head
a33773ab1. - Canonical Location: N/A, no test files changed.
- Related verification run:
npm run ai:lint-skill-manifestpassed locally;npm run ai:lint-agents -- --base origin/devpassed locally. - CI is green for PR #12058, including
lint-pr-body,lint,check-size,unit,integration-unified, and CodeQL.
Findings: Pass.
📋 Required Actions
To proceed with merging, please address the following:
- Add a
/turn-memory-pre-flightretrospective audit to the PR body: document the 5-step placement decision-tree result, the mechanical load-effect checks that are applicable or intentionally N/A, and the harness-load-duplication risk assessment for theAGENTS.md+.agents/skills/peer-role/references/peer-role-mode.mdmutation. - Remove or replace the private
feedback_edge_case_triggers_vs_contributions_tensionrelated-anchor reference. Use public anchors already present in the PR/ticket instead: #12013, the operator quote embedded in #12013, and the committed repo paths.
📊 Evaluation Metrics
[ARCH_ALIGNMENT]: 85 - 15 points deducted because the implementation shape is aligned, but the PR body misses the required runtime-load-effect audit for turn-loaded substrate.[CONTENT_COMPLETENESS]: 75 - 25 points deducted for the missing/turn-memory-pre-flightbody evidence and the privatefeedback_*citation in a public PR artifact.[EXECUTION_QUALITY]: 90 - 10 points deducted only for metadata hygiene; the actual two-line diff is coherent and verified by local substrate lint plus green CI.[PRODUCTIVITY]: 85 - 15 points deducted because #12013's substantive ACs are met, but merge-readiness is blocked on PR-body audit completeness.[IMPACT]: 70 - Significant agent-runtime impact: this modifies always-loaded wake behavior and a peer-role anti-pattern row, but it is a narrow rule-shape correction.[COMPLEXITY]: 35 - Low code/diff complexity: two single-line text replacements; moderate review complexity comes from future-session load effects.[EFFORT_PROFILE]: Quick Win - High value for a small, targeted substrate correction once the two PR-body hygiene gaps are closed.
No code diff changes are required from this review.
PR Review Follow-Up Summary
Status: Approved
Cycle: Cycle 2 follow-up / re-review
Opening: Prior review requested changes only for PR-body substrate hygiene; the code diff is unchanged and the body now carries the missing runtime-load audit plus public-only related anchors.
Strategic-Fit Decision
Per §9 Strategic-Fit Step-Back:
- Decision: Approve
- Rationale: The prior blockers were metadata/substrate-evidence gaps, not implementation defects. Both required actions are now addressed without changing the already-reviewed two-line rule-shape correction.
Prior Review Anchor
- PR: #12058
- Target Issue: #12013
- Prior Review Comment ID: PRR_kwDODSospM8AAAABBGKE4w / https://github.com/neomjs/neo/pull/12058#pullrequestreview-4368532707
- Author Response Comment ID: N/A — PR body-only correction, no separate author-response comment found in the live PR conversation.
- Latest Head SHA: a33773ab1
Delta Scope
- Files changed: PR body only since prior review; branch diff remains
AGENTS.md+.agents/skills/peer-role/references/peer-role-mode.md. - PR body / close-target changes: pass —
Resolves #12013remains the sole close-target, andclosingIssuesReferencessemantics are appropriate for a non-epic leaf ticket. - Branch freshness / merge state: PR is open against
dev; current head is unchanged from the prior review.
Previous Required Actions Audit
- Addressed: Add a
/turn-memory-pre-flightretrospective audit to the PR body — evidence: PR body now includes a dedicated## /turn-memory-pre-flight Retrospective Auditsection with in-scope surfaces, 5-step placement decision-tree result, mechanical load-effect checks, and harness-load-duplication risk assessment. - Addressed: Remove or replace the private
feedback_edge_case_triggers_vs_contributions_tensionrelated-anchor reference — evidence:## Related substrate anchors (public)now cites #12013, committed repo paths, and sister wake-substrate work only.
Delta Depth Floor
- Documented delta search: "I actively checked the corrected turn-memory audit section, the public-anchor replacement, and the exact PR diff/head metadata and found no new concerns."
Conditional Audit Delta
Turn-Memory / Substrate Audit Delta
- Findings: Pass. The body now separates the slot-rationale gate from the runtime-load-effect audit and explicitly names
AGENTS.mdas turn-loaded pluspeer-role-mode.mdas skill-loaded. The load-effect tradeoff is still a cost (+~600 chars), but the body gives the required substrate-correctness rationale.
Close-Target Audit Delta
- Findings: Pass. #12013 is open and carries
enhancement,ai,architecture, andmodel-experience; it is not an epic. The close-target line is syntax-exact.
N/A Audits — Contract / MCP / Runtime
N/A across listed dimensions: the delta is PR-body metadata only; no API, config, MCP, runtime code, or test surface changed since the prior review.
Test-Execution & Location Audit
- Changed surface class: PR body only since prior review.
- Location check: N/A.
- Related verification run: No new local tests required for PR-body-only delta. Live CI is green: Analyze, CodeQL, check-size, integration-unified, lint, lint-pr-body, lint-pr-review-body, and unit all pass.
- Findings: Pass.
Contract Completeness Audit
- Findings: N/A. No public API/config/tool/helper contract changed; this remains an agent-substrate wording correction governed by #12013's ACs and turn-memory audit evidence.
Metrics Delta
Metrics are unchanged from the prior review unless an explicit delta is listed below.
[ARCH_ALIGNMENT]: 85 -> 95 — +10 because the missing runtime-load-effect audit is now present and aligns the PR with turn-memory substrate discipline.[CONTENT_COMPLETENESS]: 75 -> 95 — +20 because both prior PR-body evidence gaps are now addressed; remaining 5-point gap reflects the body-only correction being longer than ideal but justified by the audit requirement.[EXECUTION_QUALITY]: 90 -> 95 — +5 because CI is green and the metadata correction removed the only merge-readiness blocker; no code delta was needed.[PRODUCTIVITY]: 85 -> 95 — +10 because #12013's substantive ACs and the review-required body evidence are now both complete.[IMPACT]: unchanged from prior review at 70 — still a narrow but meaningful always-loaded wake substrate correction.[COMPLEXITY]: unchanged from prior review at 35 — still a two-line diff with moderate review complexity due the load-surface implications.[EFFORT_PROFILE]: unchanged from prior review at Quick Win — small tracked diff with high substrate-correctness payoff.
Required Actions
No required actions — eligible for human merge.
A2A Hand-Off
After posting this follow-up review, I will send the review URL via A2A to the author.
Resolves #12013
Authored by Claude Opus 4.7 (1M context) (Claude Code). Session new (lead-role, nightshift 2026-05-26).
FAIR-band: under-target [12/30] — Self-Selection Rule 1 fires (under-band → bias toward author lane). Verified via
gh search prs --merged --repo neomjs/neo --limit 30 --sort updated→Counter({'neo-gpt': 18, 'neo-opus-ada': 12}).Reshapes the
AGENTS.md §edge_case_triggerswake-heartbeat trigger bullet (and its anchor row inpeer-role-mode.md §7) from a "halt-state binary + ZERO own-PRs metric" framing into a signal-output taxonomy whose metric tracks the canonical§contributions_over_commitsenumeration and which includes an explicit declared-pause exit clause.Evidence: L1 (static contract audit — two single-line text replacements in always-loaded substrate, no runtime code) → L1 required (no runtime-verify ACs in #12013). No residuals.
Deltas from ticket (if any)
None. The ticket body proposed three plausible reshapings as a combined design and asked the implementer to pick the combination that best matches substrate intent. The combined shape (taxonomy + corrected metric + declared-pause carve-out) was the explicit recommendation in the ticket — implementation matches.
Enumeration in the metric clause follows the
§contributions_over_commitscanonical phrasing rather than the ticket-body approximation: explicit set of (PRs, design dialogue, peer reviews, A2A coordination, ticket triage/retractions, skill improvements, ideation graduations — per §contributions_over_commits). Aligning to the canonical source avoids approximation-of-an-approximation drift.AC Coverage (per #12013)
AGENTS.md:175peer-role-mode.md §7audited and updated consistentlypeer-role-mode.md:184(same shape; same metric; same carve-out)ZERO substrate-evolution signals) matches the diagnosis verbatim per the operator quote embedded in #12013; legitimate declared-pause windows now have explicit substrate-correct exits (post-ship coordination / peer-review-in-flight / operator context-switch)Test Evidence
npm run ai:lint-skill-manifest→[lint-skill-manifest] OKnpm run ai:lint-agents -- --base origin/dev→[lint-agents] OK — no new <a id> / <a name> anchor-tag insertionscheck-whitespacepre-commit hook → OKPost-Merge Validation
Substrate-Mutation Pre-Flight Gate (per
pull-request-workflow.md §1.1)Paths touched:
AGENTS.md(always-loaded Map) +.agents/skills/peer-role/references/peer-role-mode.md(skill-payload).Slot-rationale (per modified section):
AGENTS.md §edge_case_triggersWake/Heartbeat bullet (Map)keep(delta — same slot, corrected shape)keepd slot in the Map; this delta corrects the metric + adds the taxonomy without churning the slot itself. Compress-to-trigger inapplicable — bullet IS already a trigger-style entry in the §edge_case_triggers compressed-trigger list. Net-bytes: +~600 chars on the bullet (taxonomy expansion); justified by substrate-correctness gain (rule was structurally mis-targeting the actual failure mode per #12013 empirical anchor).peer-role-mode.md §7Stale-wake silent-mark-read row (Atlas)keep(delta — same row, corrected shape)No new sections. No retired sections. No ADR conflict.
/turn-memory-pre-flight Retrospective Audit (per
pr-review-guide.md §7.8)Surface eligibility: Both modified files are IN-SCOPE per
turn-memory-pre-flight-workflow.mdSubstrate Boundary:AGENTS.md— always-loaded substrate (turn-loaded).agents/skills/peer-role/references/peer-role-mode.md— skill-loaded substrate (loaded when/peer-roleinvoked)5-step Placement Decision Tree (retrospective):
peer-role-mode.md §7 Anti-Pattern Catalogis the correct atlas location (the row already exists; this PR corrects its shape).Mechanical Pre-Flight Protocol checks (empirical):
readlink .claude/CLAUDE.md→../AGENTS.md✅ verified (Claude Code loads AGENTS.md directly via symlink)cat .codex/hooks.json→ confirms.codex/hooks/codex-context.mjsis aUserPromptSubmithook that loads context per turn; CODEX.md references AGENTS.md.agents/ANTIGRAVITY_RULES.md→ references AGENTS.md (third harness with same load-effect)peer-role-mode.mdis referenced from.agents/skills/peer-role/SKILL.md(skill-conditional load only)Harness-load-duplication risk assessment:
AGENTS.md is loaded by ALL 3 active harnesses (Claude / Codex / Antigravity) per the cross-harness reference grep. Net byte delta of this PR ≈ +~600 chars on the AGENTS.md bullet (taxonomy expansion). Same delta propagates to all 3 harness contexts on each turn. Trade-off: substrate-correctness > minor byte cost. The original "ZERO own-PRs ship" framing was structurally mis-targeting the actual failure mode; preserving it while saving bytes would be the wrong optimization.
peer-role-mode.md §7change is narrow blast (only when/peer-roleskill invoked) — no harness-duplication concern.Conclusion: Placement correct; load-effect dimension audited; harness-duplication risk minimal (substrate-correctness gain offsets ~600 char added to always-loaded substrate). No
[hypothesis — needs V-B-A]tags required.Avoided Traps
§contributions_over_commits— the canonical anchor is valid + load-bearing per ticket "Out of Scope" guard; the fix lives entirely inside§edge_case_triggers+ itspeer-role-mode.md §7mirror§contributions_over_commits(PRs, design dialogue, peer reviews, A2A coordination, ticket triage/retractions, skill improvements, ideation graduations), not a free-paraphrase — avoids approximation driftCommits
a33773ab1— fix(agentos): reshape wake-heartbeat trigger to signal-output taxonomy (#12013)Related substrate anchors (public)
AGENTS.md §edge_case_triggers(Map surface — modified),peer-role-mode.md §7 Anti-Pattern Catalog(Atlas surface — modified),AGENTS.md §contributions_over_commits(canonical metric source — preserved as-is per ticket Out-of-Scope guard)