메뉴
This skill should be used when the user asks to "show audit trail", "why did we make this decision", "what happened after this", "trace this decision", "show decision history", "search audit log", or needs to understand the decision history and rationale behind code changes verified through coherence check.
This skill should be used when the user asks to "define acceptance criteria", "what are the success criteria", "set quality gates", "establish acceptance tests", "define what success looks like", or needs to specify pre-declared success criteria before code execution begins.
This skill should be used when the user asks to "check this code", "validate this", "verify this implementation", "is this correct", "review this code", "check for errors", "multi-agent verification", or mentions production-critical code, financial calculations, security implementations, or high-stakes operations. Provides comprehensive multi-agent verification workflow with specialized critic agents.
This skill should be used when the user asks to "review the plan", "check the execution plan", "validate this plan", "is this plan good", or automatically after the planner agent creates an execution plan. Validates plans for completeness, clarity, measurability, and proper acceptance criteria before execution begins.
This skill should be used when the user asks to "verify critic independence", "check for overlapping failures", "validate swiss cheese model", "are the critics independent", "measure orthogonality", or needs to verify that multiple critic layers have misaligned failure modes as required by the swiss cheese model.
This skill should be used when the user asks to "use backward design", "create curriculum using UbD", "design learning outcomes first", "start with desired results", "plan assessments before activities", "design assessments first", "create outcomes-based curriculum", or references Understanding by Design framework. Also applies when user mentions aligning outcomes with assessments or avoiding activity-first planning. Provides comprehensive guidance on applying backward design methodology to create effective curricula for 1-2 day intensive workshops.
This skill should be used when the user asks to "write learning objectives", "create measurable learning outcomes", "select action verbs for objectives", "what cognitive levels should I target", "how to scaffold learning", "write course objectives", "make objectives more specific", or references Bloom's taxonomy. Also relevant when users say "my objectives aren't measurable", "how do I assess what students learned", or "progression from basic to advanced skills". Provides comprehensive guidance on applying Bloom's taxonomy to create measurable, appropriately-leveled learning objectives for intensive workshops.
| name | Audit Trail |
| description | This skill should be used when the user asks to "show audit trail", "why did we make this decision", "what happened after this", "trace this decision", "show decision history", "search audit log", or needs to understand the decision history and rationale behind code changes verified through coherence check. |
Provide bidirectional traversal of decision history, enabling users to understand why decisions were made, what happened after them, and search for specific patterns across verification sessions.
Forward: "What happened after this decision?" Backward: "Why did we make this decision?" Search: "Show all security critic rejections"
.claude/coherence-audit/ directory/audit-trail show session-abc123
Returns complete decision history for verification session.
/audit-trail search security critic rejections
Finds all sessions where security critic vetoed.
/audit-trail why was token storage changed to httpOnly
Traces decision back through critic evaluations to original requirement.
/audit-trail what files depend on auth/middleware.ts
Shows downstream impacts of changes (exposure analysis).
{
"session_id": "session-abc123",
"timestamp": "2026-02-03T14:30:00Z",
"user": "reggie",
"operation": "coherence-check",
"scope": {
"files": ["src/auth/login.ts"],
"operation": "refactor to JWT"
},
"plan": {
"acceptance_criteria": [
"All tests pass",
"No OWASP violations",
"httpOnly cookies used"
],
"critics": ["code", "security", "domain"]
},
"iterations": [
{
"iteration": 1,
"code_critic": {"verdict": "approve", "reason": "..."},
"security_critic": {"verdict": "veto", "reason": "XSS in error msgs"},
"domain_critic": {"verdict": "approve", "reason": "..."}
},
{
"iteration": 2,
"code_critic": {"verdict": "approve", "reason": "..."},
"security_critic": {"verdict": "approve", "reason": "Fixed XSS"},
"domain_critic": {"verdict": "veto", "reason": "Missing audit log"}
},
{
"iteration": 3,
"all_critics": "approve"
}
],
"final_status": "approved",
"cost": {"tokens": 15420, "overhead_pct": 38.6},
"files_changed": ["src/auth/login.ts", "src/auth/middleware.ts"]
}
Path: .claude/coherence-audit/{session_id}.json
Created after each verification session with complete decision history.
Format:
feat(auth): refactor to JWT tokens
Verified by Code Coherence (session-abc123)
- 3 iterations required
- Security critic caught XSS vulnerability
- Domain critic ensured audit trail present
Critics: code✓ security✓ domain✓
Files: src/auth/login.ts src/auth/middleware.ts
Path: .claude/coherence-audit/index.json
Maintains searchable index of all sessions for fast querying.
/audit-trail search "security critic rejections"
/audit-trail search "changes to src/auth/**"
/audit-trail search "sessions from 2026-02-01 to 2026-02-03"
/audit-trail search "sessions with >4 iterations"
When cited facts are corrected, identify downstream impacts:
/audit-trail exposure analyze "token storage decision"
Returns all sessions that depend on the token storage approach, enabling impact assessment of changes.
references/traversal-examples.md - Detailed query examplesreferences/audit-schema.md - Complete audit log JSON schema