AGENTS.md Generator
AGENTS.md is a plain-markdown file at your repo root — an open convention read by Codex, Claude Code, Cursor, and other coding agents — with direct operating instructions: setup commands, code style, testing, and PR conventions. This tool builds one from a form, section by section, free and entirely in your browser.
Why structure matters for agent instructions
Coding agents read AGENTS.md top-to-bottom as literal operating instructions, not narrative context — a wall of unstructured prose gets skimmed the same way a human would skim it. Clear H2 sections per concern (setup, style, testing) make each instruction easy for the agent to find and apply exactly when it's relevant.
Worked example
Two sections — Setup commands and Testing instructions — render as:
# ManifestKit
A static Astro site of free generators for AI-agent config files.
## Setup commands
- `npm install`
- `npm run dev`
## Testing instructions
Run `npm test` before committing. All tests must pass.FAQ
What is AGENTS.md?
AGENTS.md (agents.md) is an open, framework-agnostic convention for a plain-markdown file at your repo root that tells coding agents — Codex, Claude Code, Cursor, and others — how to work in your project: setup commands, code style, testing, and PR instructions.
How is it different from a normal README?
A README is written for humans skimming for context. AGENTS.md is written as direct operating instructions for an agent: exact commands to run, not prose explaining the project.
What sections should I include?
The most useful ones are Setup commands, Code style, Testing instructions, and PR instructions — but any section that changes how an agent should behave in your repo is worth adding.
Can I have more than one AGENTS.md?
Yes — nested AGENTS.md files in subdirectories are supported by several agents and override the root file for that subtree, useful in monorepos.