
AGENTS.md is a simple, open format designed to guide coding agents, utilized by over 60,000 open-source projects. It serves as a dedicated and predictable resource for providing context and instructions that assist AI coding agents in effectively working on a project. Unlike traditional README.md files, which are tailored for human contributors, AGENTS.md focuses on delivering the detailed context that coding agents require, such as build steps, tests, and conventions that may not be relevant to human contributors.
This format emerged from collaborative efforts within the AI software development ecosystem and is now stewarded by the Agentic AI Foundation under the Linux Foundation. AGENTS.md aims to maintain and evolve as an open format that benefits the entire developer community, regardless of the coding agent in use.
Provides a clear, predictable place for coding agents to find instructions.
Keeps README files concise and focused on human contributors.
Allows for tailored instructions in large monorepos through nested AGENTS.md files.
AGENTS.md is a simple, open format designed to guide coding agents, serving as a dedicated and predictable resource for providing context and instructions necessary for AI coding agents to effectively work on projects. It complements traditional README.md files, which are primarily aimed at human contributors, by containing detailed information that coding agents require, such as build steps, tests, and coding conventions. This separation ensures that README files remain concise and focused on human users while giving agents a clear place for their instructions.
Some of the key features and capabilities of AGENTS.md include:
Standard Markdown format, allowing flexibility in structuring content.
Ability to list testing commands that agents can execute automatically.
Support for nested AGENTS.md files in large monorepos, enabling tailored instructions for subprojects.
Living documentation that can be updated as project needs evolve.
Facilitates the migration of existing documentation by renaming files to AGENTS.md.
AGENTS.md provides a dedicated and predictable format for guiding coding agents, which is beneficial for over 60,000 open-source projects. By serving as a README specifically for agents, it allows developers to include detailed context and instructions that are essential for AI coding agents to effectively work on projects. This separation from traditional README.md files ensures that the latter remains concise and focused on human contributors, while AGENTS.md caters to the specific needs of coding agents.
The advantages of using AGENTS.md include:
Clear and predictable instructions for coding agents.
Ability to include detailed build steps, tests, and conventions without cluttering the README.
Support for nested AGENTS.md files in large monorepos, allowing tailored instructions for subprojects.
Compatibility with a growing ecosystem of AI coding agents and tools.
To get started with AGENTS.md, first create an AGENTS.md file at the root of your repository. Most coding agents can even scaffold one for you if you ask nicely. This file serves as a dedicated space for providing context and instructions that help AI coding agents work effectively on your project.
Next, cover what matters by adding sections that assist the agent in understanding your project. Popular choices include:
Project overview
Build and test commands
Code style guidelines
Testing instructions
Security considerations
Ready to see what AGENTS.md can do for you?and experience the benefits firsthand.
Navigate to the tool's official website.