generate-agents-md

Generate AGENTS.md file for a codebase to help AI agents understand the project structure, build commands, testing, and conventions. Use when the user asks to create an AGENTS.md, initialize agent documentation, or analyze a codebase for agent context. Also use when setting up a new repository for AI-assisted development.

Generate AGENTS.md

Generate an AGENTS.md file containing essential information for AI agents working in this repository.

What to Include

The AGENTS.md file should be ~20 lines and contain:

Build/lint/test commands - Especially how to run a single test
Architecture overview - Important subprojects, internal APIs, databases
Code style guidelines - Imports, conventions, formatting, types, naming, error handling

Process

First check if AGENTS.md or AGENT.md already exists - if so, update it instead of overwriting
Check for existing rules files to incorporate:
- .cursor/rules/ or .cursorrules
- CLAUDE.md
- .windsurfrules
- .clinerules
- .goosehints
- .github/copilot-instructions.md
Analyze the codebase structure:
- Look at package.json, Cargo.toml, pyproject.toml, etc.
- Identify test frameworks and commands
- Find lint/format configuration files
- Map out the directory structure
Write concise AGENTS.md (~20 lines)

Output

Save the generated file as AGENTS.md in the repository root.

Example Structure

# Agent Guidelines

## Commands
- Build: `npm run build`
- Test: `npm test` (single: `npm test -- file.test.ts`)
- Lint: `npm run lint`

## Architecture
- Monorepo with packages in `/packages/`
- Main API in `/packages/api/`
- Database: PostgreSQL with Prisma

## Style
- TypeScript with strict mode
- ESLint + Prettier
- Named exports preferred
- Use `logger` from `@/utils` not console