renatocaliari/cali-post-execution-check

Post-Execution Verification

Systematically verify that planned work was executed correctly, completely, and safely.

When to Activate

• After completing any multi-step task
• After executing a plan or spec
• After implementing a feature
• After refactoring code
• When user says "done", "finished", "completed"
• When user asks "check my work", "verify", "gap analysis"

Verification Workflow

Phase 1: Plan Compliance (10 min)

1. Read the plan/spec — Find the original plan, spec, or task list
2. Check each item — For each planned item:
   • ✅ Was it implemented?
   • ⚠️ Was it implemented correctly?
   • ❌ Was it skipped?
   • 🔄 Was it changed from the plan?
3. Output: Gap analysis table with status per item

| Plan Item | Status | Notes |
|-----------|--------|-------|
| Create skill X | ✅ Done | — |
| Update config Y | ⚠️ Partial | Missing validation |
| Delete old file Z | ❌ Skipped | Still exists |

Phase 2: Implementation Quality (5 min)

1. Check for obvious issues:
   • Syntax errors
   • Missing imports
   • Broken references
   • Inconsistent naming
2. Check for anti-patterns:
   • Secrets in code
   • Global mutable state
   • fmt.Sprintf with HTML
   • God functions (>100 lines)
3. Output: Quality issues list

Phase 3: Edge Cases (5 min)

1. Think about failure modes:
   • What happens if input is null/empty?
   • What happens if network fails?
   • What happens if file doesn't exist?
   • What happens if permission denied?
2. Think about concurrency:
   • Race conditions?
   • Deadlocks?
   • State corruption?
3. Output: Edge case list

Phase 4: Documentation & Tests (5 min)

1. Identify affected documentation:
   • README.md — needs update?
   • AGENTS.md — needs update?
   • CHANGELOG.md — needs entry?
   • Skill SKILL.md — needs update?
   • API docs — needs update?
2. Identify affected tests:
   • Unit tests — need new/updated tests?
   • Integration tests — need updates?
   • Manual test cases — need new cases?
3. Output: Documentation/test update list

Phase 5: Summary Report (2 min)

Compile all findings into a structured report:

## Post-Execution Report

### Plan Compliance: X/Y items complete
- ✅ Completed: N items
- ⚠️ Partial: N items
- ❌ Skipped: N items

### Quality Issues: N found
- Critical: N
- Warning: N
- Info: N

### Edge Cases: N identified
- [List edge cases]

### Documentation Updates Needed: N
- [List documentation]

### Test Updates Needed: N
- [List tests]

Output Format

Always produce a structured report. Use the format above.

Integration with Other Skills

| Skill | When to Use | |---|---| | cali-skill-validator | After creating/modifying skills | | cali-agents-md-validator | After modifying AGENTS.md | | cali-product-testing-execution | After implementing features | | dogfood | After UI changes |

Warnings

• Don't skip phases — Each phase catches different issues
• Don't assume — Verify, don't guess
• Don't rush — 20 minutes of checking saves hours of debugging
• Don't ignore warnings — Even minor issues compound