| name | react-quality-engine |
| description | Meta-skill: React Quality Orchestrator for AFENDA-HYBRID. Coordinates architecture, performance, accessibility, and design-pattern skills for audits, scoring (Component Quality Index), and confidence-tiered refactors. Use for full-route reviews, legacy refactors, and governance checks. |
| metadata | {"author":"AFENDA team","version":"2026.03.23","source":"AFENDA-HYBRID governance + React docs","official_docs":"https://react.dev"} |
This skill coordinates:
Use AGENTS.md for routing (Skill Invocation Matrix) and repo-wide non-goals.
Automated package-layer gate: pnpm react-governance (see tools/react-governance/README.md; check --json reserves summary.score for scoring pipelines).
Provenance: see GENERATION.md.
Orchestration protocol
- Classify the task (audit vs targeted fix vs refactor).
- Pick primary skill from the matrix in
AGENTS.md; pull secondaries only as needed.
- For audits, walk dimensions in scoring-model and record evidence.
- For code changes, assign a confidence tier per refactor-safety-rules.
- Output for humans: problem, why it matters, suggested fix, before/after (when tier allows), tier, and score impact.
Core references
Non-Goals
This skill does not:
- Bypass human review for tier C/D changes
- Change product requirements or authorization rules
- Run production deploys or mutate infrastructure
- Implement CI jobs or ESLint rule packs (Phase 2 tooling)
AFENDA Monorepo Reality Mapping
| Area | Enforce |
|---|
| All React packages | Scoring and tiers apply; evidence should cite file paths and pattern docs |
| erp-view-pack | Cross-check _CONVENTIONS.md / _AUDIT_MATRIX.md when scoring pattern compliance |
| apps/web | Next.js boundaries and server actions weigh heavily under architecture + pattern consistency |
| db / non-UI | Out of scope except type safety or shared types touching UI |
Entry point
Repository agents should start at AGENTS.md, then open this meta-skill when performing quality audits or guided refactors.
