An AGENTS.md tells a coding agent how your project works: the commands it may run, the conventions it must follow, and the files it must not touch. These 12 templates are complete files, not outlines. Pick the closest project type, adapt the commands to your repo, and commit it to the root.
App Router SaaS with Supabase auth, Postgres, and Stripe billing.
86 lines
Multi-package workspace with shared UI, config, and typed contracts.
105 lines
Async FastAPI service with SQLAlchemy, Alembic, and pytest.
86 lines
Django app with per-app structure, Celery jobs, and Stripe billing.
84 lines
Expo Router app with typed navigation, native modules, and EAS builds.
96 lines
Flutter app with Riverpod state, go_router navigation, and codegen.
92 lines
Laravel app with Eloquent, queues, Cashier billing, and Pest tests.
91 lines
SvelteKit app with server load functions, form actions, and Vitest.
89 lines
Published npm package with a stable API, changesets, and docs.
91 lines
Shared codebase serving several client projects with isolated config.
99 lines
Training and inference repo with tracked experiments and versioned data.
98 lines
A short, stack-agnostic AGENTS.md to fill in and extend.
79 lines
Claude Code reads CLAUDE.md. Most other agents and editors have converged on AGENTS.md. The content is the same, so the working pattern is one real file plus a symlink, which keeps the rules from drifting apart:
ln -s AGENTS.md CLAUDE.md Full reasoning, including the Windows caveat, is in CLAUDE.md vs AGENTS.md. Looking for boilerplates that already ship these files? See agent-ready boilerplates.