CLI that generates an AGENTS.md for your project. It reads your existing config files - package.json, pyproject.toml, Cargo.toml, go.mod, and others - and produces a ready-to-fill template with detected commands, linters, and project structure.
AGENTS.md is the file AI coding agents (Claude Code, Codex, Cursor, Gemini CLI, and others) read to understand how to work in a repo.
npx @horiastanxd/agentsmdOr install globally:
npm install -g @horiastanxd/agentsmd
agentsmdBy default it writes AGENTS.md in the current directory. Pass a path to run it elsewhere:
agentsmd ./my-project| Flag | Description |
|---|---|
--output <file> |
Write to a different file name |
--stdout |
Print to stdout instead of writing |
--overwrite |
Overwrite an existing AGENTS.md |
--version |
Print version |
--help |
Print help |
| Ecosystem | Files read | Extracted |
|---|---|---|
| Node.js | package.json, lock files |
scripts, package manager, framework, linters |
| Python | pyproject.toml, requirements.txt, Pipfile, poetry.lock, uv.lock |
package manager, test runner, linters |
| Rust | Cargo.toml |
build/test/lint commands, workspace flag |
| Go | go.mod, .golangci.yml |
module name, build/test/lint commands |
| Ruby | Gemfile |
bundler, rspec/rake, RuboCop |
| Java/Kotlin | pom.xml, build.gradle |
Maven or Gradle commands |
| PHP | composer.json |
composer install, PHPUnit |
For multi-language monorepos, sections for each detected ecosystem are included.
Running agentsmd --stdout on a typical Next.js project:
# AGENTS.md
Instructions for AI coding agents working in this repository.
## Setup
npm install
## Commands
- **build**: `npm run build`
- **dev**: `npm run dev`
- **test**: `npm run test`
- **lint**: `npm run lint`
## Code style
Linters / formatters in use: ESLint, Prettier, TypeScript
## Architecture
Framework: Next.js
_Fill in component boundaries, data flow, and any non-obvious constraints here._
## Conventions
_Add project-specific conventions: naming, file organisation, commit format, etc._
Most repos that work with AI agents need some form of AGENTS.md (or CLAUDE.md, .cursorrules, etc.) to give the agent context on how to build, test, and lint the code. Writing one from scratch every time is tedious. This tool does the mechanical part - scanning configs - and leaves the judgment calls (architecture, conventions, gotchas) for you to fill in.
Node.js 18 or later. No runtime dependencies.
MIT