cursor-rules-collection

Battle-tested rules for AI coding agents. Every rule has been verified to actually change agent output — if a rule didn't produce measurably different code in testing, it's not here.

What's Included

| Skill | Impact | What It Covers | |-------|--------|----------------| | react-patterns | HIGH | Hooks, composition, memoization, Server Components | | nextjs-app-router | HIGH | App Router, Server Actions, caching, revalidation | | typescript-strict | HIGH | Strict mode patterns, branded types, discriminated unions | | express-api | HIGH | Middleware, error handling, validation, security headers | | python-backend | HIGH | Type hints, async patterns, Pydantic, structured errors | | error-handling | CRITICAL | Error boundaries, recovery, user-facing messages | | security | CRITICAL | Auth, injection prevention, XSS, CSRF, secrets | | testing-patterns | HIGH | Test structure, mocking, coverage, edge cases | | api-design | MEDIUM | REST conventions, validation, versioning, pagination | | performance | HIGH | Lazy loading, caching, bundle size, code splitting |

When to Apply

Use these rules when:

• Writing new components, APIs, or backend services
• Refactoring existing code for better patterns
• Reviewing code for common mistakes
• Setting up new projects with good defaults

What Makes These Different

Every rule contains something you can't get from a generic LLM prompt:

• Version-specific gotchas (e.g., Express 5 vs 4 patterns)
• Tested edge cases with before/after examples
• Anti-patterns with explanations of why they fail
• Framework-specific patterns, not language-generic advice

Related Tools

• cursor-doctor — lint your rules, detect conflicts, see load order
• Full collection — 98 rules across 30+ frameworks