| name | logos-review |
| description | Parallel Agent fan-out for Logos code quality review. Auto-run when: preparing a release, before a demo, after large refactors across Logos, or user asks for a code quality review. Invoke proactively without asking. |
Full Logos code quality review using parallel Agent fan-out. Covers all three layers.
Launch 3 parallel Agent calls (subagent_type: Explore, thoroughness: very thorough):
Agent 1 — API layer (logos/api/, logos/routes/):
- Endpoint coverage and consistency
- Error handling and response schemas
- Input validation
- Performance concerns (N+1 queries, blocking calls)
Agent 2 — Engine layer (logos/engine/, logos/rules/):
- Rule correctness and phase ordering
- Error recovery and edge cases
- Reactive cascade behavior
- Resource cleanup
Agent 3 — Frontend (hapax-logos/src/):
- Component quality and prop interfaces
- Accessibility
- Performance (re-renders, bundle size)
- Design language compliance (check against
docs/logos-design-language.md)
Synthesize into a go/no-go report with:
- Critical issues (must fix)
- Warnings (should fix)
- Observations (nice to fix)
- Overall quality assessment
