\r\n Your AI agents are only as good as their instructions.\r\n
\r\n","readmeFilename":"README.md","_rev":"1-f4ebe86429ed5efc82cc7d4de03bc4dd"}\r\n
One CLI to rule all AI coding agents.
\r\nAuto-generate, sync, and lint rules for Claude Code, Cursor, Windsurf, Copilot, and Gemini.
\r\n\r\n\r\n\r\n Install •\r\n Quick Start •\r\n Commands •\r\n Rule Packs •\r\n Agents\r\n
\r\n\r\n---\r\n\r\n## The Problem\r\n\r\nYou use AI coding agents. Maybe Claude Code, maybe Cursor, maybe both. Each has its own rules file:\r\n\r\n| Agent | File |\r\n|-------|------|\r\n| Claude Code | `CLAUDE.md` |\r\n| Cursor | `.cursorrules` |\r\n| Windsurf | `.windsurfrules` |\r\n| GitHub Copilot | `.github/copilot-instructions.md` |\r\n| Gemini | `.gemini/settings.json` |\r\n\r\nThey're all slightly different. They get out of sync. They get bloated until the AI ignores them. You wrote them by hand and half the rules are too vague to be useful.\r\n\r\n## The Solution\r\n\r\n```bash\r\nnpx agentrules init\r\n```\r\n\r\nScans your codebase, detects your stack, generates rules for all your agents. One command.\r\n\r\n```\r\nDetected: Next.js 15 + TypeScript + Tailwind + Prisma + Vitest\r\n\r\nGenerated 24 rules from your stack:\r\n 12 from detected frameworks\r\n 8 from @agentrules/nextjs pack\r\n 4 from @agentrules/typescript pack\r\n\r\nWrote:\r\n CLAUDE.md (24 rules)\r\n .cursorrules (24 rules)\r\n .github/copilot-instructions.md (24 rules)\r\n```\r\n\r\n## Install\r\n\r\n```bash\r\nnpm install -g agentrules\r\n# or use directly\r\nnpx agentrules init\r\n```\r\n\r\n## Quick Start\r\n\r\n```bash\r\n# 1. Generate rules from your codebase\r\nnpx agentrules init\r\n\r\n# 2. Check rule quality\r\nnpx agentrules lint\r\n\r\n# 3. Add community rule packs\r\nnpx agentrules add nextjs\r\nnpx agentrules add tailwind\r\n\r\n# 4. Sync to all agents\r\nnpx agentrules sync\r\n```\r\n\r\n## Commands\r\n\r\n### `agentrules init`\r\nScans your project and generates rules for all detected AI agents.\r\n\r\n```bash\r\nagentrules init # Auto-detect everything\r\nagentrules init --agents claude,cursor # Only generate for specific agents\r\nagentrules init --force # Overwrite existing rules\r\n```\r\n\r\n**Detects:**\r\n- Framework (Next.js, React, Vue, Svelte, Express, FastAPI, Django...)\r\n- Language (TypeScript, JavaScript, Python, Go, Rust...)\r\n- Package manager (npm, yarn, pnpm, bun, pip, cargo...)\r\n- Testing (Jest, Vitest, Pytest, Go test...)\r\n- Linter/formatter (ESLint, Prettier, Biome, Black, Ruff...)\r\n- Database (Prisma, Drizzle, Supabase, SQLAlchemy...)\r\n- CSS (Tailwind, CSS Modules, styled-components...)\r\n- Monorepo (Turborepo, Nx, Lerna...)\r\n\r\n### `agentrules sync`\r\nSyncs your source-of-truth rules to all agent formats.\r\n\r\n```bash\r\nagentrules sync # Sync to all agents\r\nagentrules sync --agents claude # Sync to specific agent\r\nagentrules sync --dry-run # Preview without writing\r\n```\r\n\r\nYour rules live in `.agentrules/rules.md`. Edit that file, run sync, and every agent gets the update.\r\n\r\n### `agentrules lint`\r\nCatches problems in your rules before they confuse your AI.\r\n\r\n```bash\r\nagentrules lint\r\n```\r\n\r\n**Catches:**\r\n- Rules that are too vague (\"write good code\")\r\n- Contradictory rules (\"always use semicolons\" + \"never use semicolons\")\r\n- Duplicate rules\r\n- File too long (AI agents ignore bloated rule files)\r\n- Missing rules for your detected stack\r\n\r\n```\r\n 2 errors, 3 warnings\r\n\r\n error Line 14: Rule is too vague: \"Write clean code\"\r\n Suggestion: Specify what \"clean\" means for your project\r\n\r\n error Line 23 conflicts with Line 45:\r\n \"Always use arrow functions\" vs \"Use function declarations for hoisting\"\r\n\r\n warn Rules file is 847 lines. AI agents lose accuracy above 500 lines.\r\n Consider splitting into essential vs nice-to-have rules.\r\n\r\n warn No testing rules found. Your stack uses Vitest.\r\n Run: agentrules add testing\r\n\r\n warn Duplicate rule on lines 34 and 67:\r\n Both say \"Use TypeScript strict mode\"\r\n```\r\n\r\n### `agentrules add\r\n Your AI agents are only as good as their instructions.\r\n
\r\n","readmeFilename":"README.md","_rev":"1-f4ebe86429ed5efc82cc7d4de03bc4dd"}