chore(skills): add code quality review skills and cleanup command

[waleedlatif1]

Summary

• Add 6 new skills: memo, callback, state, react-query, emcn-design-review, cleanup
• Each skill has self-contained anti-pattern checklists and codebase-specific context
• /cleanup chains all 6 skills sequentially for full code review
• All skills registered in .agents/skills, .claude/commands, .cursor/commands

Type of Change

• New feature

Testing

Tested manually

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
docs	Ready	Preview, Comment	Apr 13, 2026 6:31pm

[cursor]

PR Summary

Low Risk
Low risk because this PR only adds markdown-based skill/command definitions and a wrapper cleanup command; it doesn’t change runtime application code or behavior.

Overview
Adds five new review skills—you-might-not-need-a-memo, you-might-not-need-a-callback, you-might-not-need-state, react-query-best-practices, and emcn-design-review—each documenting the checks to perform and whether to auto-fix or only propose changes.

Introduces a cleanup skill/command that chains you-might-not-need-an-effect plus the new skills in a single sequential pass, and registers the new commands for both Claude (.claude/commands) and Cursor (.cursor/commands).

^{Reviewed by Cursor Bugbot for commit 6686a6d. Configure here.}

[greptile-apps]

Greptile Summary

This PR adds 5 new AI code-quality review skills (memo, callback, state, react-query, emcn-design-review) plus a cleanup orchestrator that chains all six skills (including the pre-existing you-might-not-need-an-effect) sequentially. Each skill is registered in three places — .agents/skills/, .claude/commands/, and .cursor/commands/ — so it works across all three AI tooling contexts. The rules in every file are consistent with the project's CLAUDE.md guidelines and architecture conventions.

Confidence Score: 5/5

Safe to merge — purely documentation/configuration additions with no executable code changes.

All 18 files are Markdown skill/command definitions. Content is accurate, consistent with CLAUDE.md, and well-structured. No P0 or P1 findings.

No files require special attention.

Important Files Changed

Filename	Overview
.agents/skills/cleanup/SKILL.md	Orchestrator skill that chains all 6 quality skills in sequence; references the pre-existing /you-might-not-need-an-effect as step 1.
.agents/skills/emcn-design-review/SKILL.md	Comprehensive EMCN design-system review skill with full token tables, component rules, and anti-pattern checklist; consistent with CLAUDE.md and emcn conventions.
.agents/skills/react-query-best-practices/SKILL.md	React Query audit skill covering key factories, staleTime, mutations, and server-state ownership; rules mirror CLAUDE.md React Query guidelines exactly.
.agents/skills/you-might-not-need-a-callback/SKILL.md	useCallback anti-pattern detector; includes codebase-specific whitelist for the useRef+empty-deps stable-callback pattern.
.agents/skills/you-might-not-need-a-memo/SKILL.md	useMemo/React.memo anti-pattern detector covering 7 common misuse patterns; clean and accurate.
.agents/skills/you-might-not-need-state/SKILL.md	useState anti-pattern skill with 6 detection rules; correctly aligns React Query / Zustand / useState ownership with CLAUDE.md architecture.

_{Reviews (1): Last reviewed commit: "chore(skills): add code quality review s..." | Re-trigger Greptile}