Agentsmd — Development skill for Claude Code

agentsmd

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.

Usage

npx @horiastanxd/agentsmd

Or install globally:

npm install -g @horiastanxd/agentsmd
agentsmd

By default it writes `AGENTS.md` in the current directory. Pass a path to run it elsewhere:

agentsmd ./my-project

Options

Flag	Description
--output	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

What it detects

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.

Output example

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 i