AGENTS.md Generator

Generate an AGENTS.md for coding agents from Node, Python, Go or monorepo presets, with setup, build, test and lint commands and done criteria.

Presets:

Project name

Project overview

Setup commands (one per line)

Build command

Test command

Lint command

Run / dev command

Code style & conventions

Commit & PR conventions

How to verify your change is done

Security & secrets — do NOT

Monorepo packages (optional)

No packages added. Use this for monorepos with per-package commands.

AGENTS.md

Structure looks complete — no lint issues.

Generated Markdown

# AGENTS.md

> Operating guide for AI coding agents working in **My App**. Read this before editing, building, or opening a PR.

## Project overview

A TypeScript application. Make small, typed, well-tested changes. Read the existing code before introducing new patterns.

## Setup

```bash
npm install
```

## Commands

Run the exact invocation below. Do not invent flags or scripts.

- **Build:** `npm run build`
- **Test:** `npm test`
- **Lint:** `npm run lint`
- **Run / dev:** `npm run dev`

## Code style & conventions

- TypeScript strict mode; avoid `any` without a clear reason.
- Prefer named exports and small, single-purpose modules.
- Format with Prettier; do not hand-format.
- Reuse existing logging and error utilities instead of `console.log`.

## Commit & PR conventions

- Use Conventional Commits (feat:, fix:, chore:, docs:).
- One logical change per PR; keep the diff small and reviewable.
- Update or add tests in the same change as the code.

## How to verify your change is done

- `npm test` passes with zero failures.
- `npm run typecheck` reports no errors.
- `npm run lint` is clean.
- The change works when run with `npm run dev`.

## Security & secrets — do NOT

- Never commit secrets, credentials, or `.env` files.
- Never log API keys, tokens, or personal data.
- Do not disable or weaken auth checks to make a test pass.

What this tool does

The AGENTS.md Generator builds a best-practice AGENTS.md file that tells coding agents how to work in your repo. It produces command-first sections for build, test, and lint, records your conventions, and writes verifiable done-criteria, so agents follow the same rules a new teammate would. It runs in your browser with no signup.

Updated October 2, 2026. Provided as is. Check the output before you rely on it in production.

How to use AGENTS.md Generator

1

Pick a preset

Start from Node/TypeScript, Python, Go, or Monorepo to pre-fill sensible commands and conventions, then edit them to match your project.
2

Enter your exact commands

Fill the setup, build, test, lint, and run fields with the real invocations an agent should run. Command-first beats prose every time.
3

Define done-criteria

Spell out how an agent verifies a change is complete: tests pass, type-check clean, lint clean. This is the highest-value section.
4

Check the lint panel

Review the warnings for empty command sections or placeholder text like TODO, and fill the gaps before shipping the file.
5

Copy or download AGENTS.md

Copy the generated Markdown or download the AGENTS.md file and drop it in your repository root.

Questions and answers

What is an AGENTS.md file?

AGENTS.md is a Markdown file at a repository root that tells coding agents how to set up, build, test and change the project. This generator assembles one from your commands, conventions and done criteria, adds per-package sections for monorepos, and lints for empty command sections and placeholders like TODO.

How does this generator work?

Fill in your setup, build, test, lint, and run commands plus your conventions, and it assembles a clean, command-first Markdown file. A built-in lint flags empty command sections and vague placeholders like TODO.

What makes a good AGENTS.md?

Exact command invocations (not prose), explicit code-style rules, and concrete 'how to verify your change is done' criteria. Agents follow commands far more reliably than vague descriptions.

Does it support monorepos?

Yes. Use the Monorepo preset or add per-package sections with their own paths and commands, so an agent knows which workspace flag to pass for each package.

Is my input sent anywhere?

No. The file is assembled entirely in your browser. Nothing you type is uploaded, logged, or stored on a server.

For AI agents: how to call this tool

Machine-readable contract, endpoints and examples. Humans can ignore this section.

Best Path For Builders

Browser workflow

Runs instantly in the browser with private local processing and copy/export-ready output.

Browser Workflow

This tool is optimized for instant in-browser execution with local data handling. Run it here and copy/export the output directly.

/agents-md-generator/

For automation planning, fetch the canonical contract at /api/tool/agents-md-generator.json.

AGENTS.md Generator