Create AGENTS.md files for project-specific inline rules. Use when adding small, project-specific instructions that should be committed in repos.
Create AGENTS.md
Guide for creating AGENTS.md files for project-specific inline rules.
When to Use AGENTS.md
- Small, project-specific instructions that should be committed in the repo
- Folder-scoped rules for specific directories
- Package-specific instructions in monorepos
- Test-specific guidance in test directories
When NOT to Use AGENTS.md
- Reusable knowledge across projects → Use skills
- Large documentation → Use skills with references
- Complex workflows → Use skills with scripts
AGENTS.md Structure
AGENTS.md is a simple markdown file without metadata:
# Project Instructions
## Code Style
- Use TypeScript for all new files
- Prefer functional components in React
- Use snake_case for database columns
## Architecture
- Follow the repository pattern
- Keep business logic in service layers
Location
- Project root:
AGENTS.md– Primary, inline instructions and references for the whole project (commands, tech stack, testing, code style, architecture, safety boundaries). - Subdirectories:
subdirectory/AGENTS.md– Folder- or package-scoped instructions when local behavior meaningfully diverges from the root (e.g., a specific package, service, or test tree). - Nested support: Agents typically combine instructions from the closest
AGENTS.mdwith parent ones; keep root general and use nestedAGENTS.mdonly where you truly need more specific rules.
Best Practices
- Keep AGENTS.md files small and focused
- Use for project-specific conventions
- Prefer short, concrete references over long prose:
- Link to project docs, specs, and runbooks
- Point to example files or directories (e.g.,
see src/api/users.ts for canonical pattern) - Include the most important commands with exact CLI invocations
- Reference existing code examples when possible
- Update as project evolves
References
For detailed best practices, see references/best-practices.md.
You Might Also Like
Related Skills

update-docs
This skill should be used when the user asks to "update documentation for my changes", "check docs for this PR", "what docs need updating", "sync docs with code", "scaffold docs for this feature", "document this feature", "review docs completeness", "add docs for this change", "what documentation is affected", "docs impact", or mentions "docs/", "docs/01-app", "docs/02-pages", "MDX", "documentation update", "API reference", ".mdx files". Provides guided workflow for updating Next.js documentation based on code changes.
vercel
docstring
Write docstrings for PyTorch functions and methods following PyTorch conventions. Use when writing or updating docstrings in PyTorch code.
pytorch
docs-writer
Always use this skill when the task involves writing, reviewing, or editing files in the `/docs` directory or any `.md` files in the repository.
google-gemini
write-concept
Write or review JavaScript concept documentation pages for the 33 JavaScript Concepts project, following strict structure and quality guidelines
leonardomso
resource-curator
Find, evaluate, and maintain high-quality external resources for JavaScript concept documentation, including auditing for broken and outdated links
leonardomso
doc-coauthoring
Guide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.
anthropics