docs(learnings): record L9 governance git/memory-lock gotchas

[cryptoxdog]

What

Append two cross-repo governance lessons to .claude/rules/learnings.md, in the file's documented - **[date] Category: description** — explanation format. Docs-only; no code or contract changes.

Why

These gotchas surfaced during L9 CI work and cost real debugging time; recording them helps future agent sessions in any governed repo (the memory-lock gate applies to CEG too):

1. The compound git add && git commit trap — the L9 memory-lock PreToolUse gate denies a compound Bash git add && git commit as a single unit, so git add never runs and only previously-staged changes commit. The result is a commit that silently diverges from the validated working tree. Fix: stage in a separate Bash call from the lock-gated commit, and verify with git show HEAD:<path> after every governed commit.
2. Lock session / state-root binding — a lock acquired via the memory_lock.py CLI must match the gate's real CLAUDE_CODE_SESSION_ID and CLAUDE_PROJECT_DIR (=/home/user); a manual run with no hook event writes session_id="unknown-session" and anchors state at the cwd fallback, so the gate ignores it. Locks also carry a short TTL — re-acquire before each governed write.

Scope

• One file changed: .claude/rules/learnings.md (+2 lines).
• Single commit above main.

---

Generated by Claude Code

[github-actions]

✅ PR reviewable size is within recommended limits

[github-actions]

L9 Audit Harness Report

• Generated: 2026-08-05T01:09:00.575010+00:00
• Repo root: /home/runner/work/Cognitive.Engine.Graphs/Cognitive.Engine.Graphs
• Overall result: ✅ PASSED
• Exit code: 0

Step Results

Step	Status	Exit Code	Notes
Architecture Audit	✅ Passed	0
Spec Coverage	✅ Passed	0
Contract Wiring	✅ Passed	0

Architecture Audit Findings

Severity	Count
🔴 CRITICAL	0
🟠 HIGH	0
🟡 MEDIUM	25
🔵 LOW	0

See artifacts/audit_report.md for full details.

Spec Coverage

• ✅ Implemented: 37
• ⚠️ Partial: 9
• ❌ Missing: 0
• Total features: 46

Category	Implemented	Partial	Missing	Total
gates	10	0	0	10
scoring	7	0	0	7
v1.1_node	2	0	0	2
v1.1_edge	2	0	0	2
v1.1_action	0	2	0	2
v1.1_scoring	1	1	0	2
action_handler	0	6	0	6
gds_algorithm	5	0	0	5
research_pattern	10	0	0	10

See artifacts/coverage_report.md for full details.

Next Steps

All checks passed. Safe to merge.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[Copilot]

Pull request overview

This PR appends two governance “learnings” entries to .claude/rules/learnings.md to capture operational gotchas around the L9 memory-lock gate, helping future agent sessions avoid subtle git/lock pitfalls.

Changes:

• Documented a git workflow pitfall where a lock-gated compound git add && git commit can result in committing only previously staged changes.
• Documented a memory-lock pitfall where manually acquired locks may be ignored if the session/state-root context doesn’t match the hook environment.

[cryptoxdog]

This belongs elsewhere in the https://github.com/Quantum-L9/Cursor-Governance learnings not .claude

[cryptoxdog]

Agreed — this is a governance-wide gate lesson, not a CEG concern. Relocating it to Quantum-L9/Cursor-Governance (the canonical lessons DB, learning/failures/repeated-mistakes.md) and backing out this misplacement. Closing this PR and removing the .claude/rules/learnings.md change from CEG.

(For the record, the failing Baseline Ratchet / Required Tests here is unrelated to this docs change: tests/test_handlers.py::test_match_returns_structure fails because handle_match's early-return path omits query_id while the full-path response includes it — a pre-existing main-side issue the ratchet surfaced after the quarantine ledger was emptied in 8be58df. Not fixing it here since this PR is being closed; happy to open a separate fix PR if useful.)

---

Generated by Claude Code