Source: https://arara.code.app.br/en/articles/agents-md-rules-for-ai-agents
Language: en-US
Published: 2026-11-16T08:00:00-03:00

Guides and comparisons

# AGENTS.md: how to guide AI coding agents

AGENTS.md is an open Markdown standard placed at the repository root to give coding agents predictable project instructions. It acts like a README for AI, unifying build commands, test suites, style rules, and guardrails in one cross-tool file.

Arara Code

Published and sources checked on November 16, 2026.

Published by the Arara Code project based on the official documentation cited.

## Why AGENTS.md was created

Before AGENTS.md, every assistant demanded its own proprietary config file—such as .cursorrules, CLAUDE.md, or custom extension settings—creating duplication and drift for teams using multiple tools.

With the open standard hosted at agents.md, one file describes how your codebase works and is natively read by tools including OpenAI Codex, Claude Code, Gemini, GitHub Copilot, and Cursor.

[AGENTS.md: official open standard specification](https://agents.md/)

## What to include in AGENTS.md

Because it has no rigid schema, the file uses plain Markdown headings: setup and dev commands, required typecheck and test scripts, architecture conventions, and explicit guardrails on files agents should never edit.

In large monorepos, you can also place nested AGENTS.md files inside subdirectories to scope package-specific rules without bloating global context.

[GitHub: official agentsmd/agents.md repository](https://github.com/agentsmd/agents.md)

Related

[MCP (Model Context Protocol): what it is and how to use itLearn what MCP (Model Context Protocol) is, how servers, tools and transports work, which security rules to follow and how to use it in Arara Code.](https://arara.code.app.br/en/articles/what-is-mcp)

## Fewer hallucinations and lower token waste

Giving agents exact verification commands and module paths upfront prevents expensive discovery loops where assistants guess package scripts.

That cuts command failures, keeps pull requests consistent, and makes automated edits far more predictable.

[AGENTS.md: official open standard specification](https://agents.md/)

## AGENTS.md inside Arara Code

Arara Code respects your workspace AGENTS.md and applies your project rules consistently whether you run Claude Code, Codex, Kiro, or Antigravity.

Every teammate keeps the same quality bar even when developers pick different models or agents inside the IDE.

[Arara Code: public IDE repository](https://github.com/arara-ide/arara-code)

## Sources and further reading

- [AGENTS.md: official open standard specification](https://agents.md/)
- [GitHub: official agentsmd/agents.md repository](https://github.com/agentsmd/agents.md)
- [Arara Code: public IDE repository](https://github.com/arara-ide/arara-code)

[Read as Markdown](https://arara.code.app.br/en/articles/agents-md-rules-for-ai-agents.md)

## Keep exploring

### [Context engineering for AI coding: how to structure](https://arara.code.app.br/en/articles/context-engineering-for-ai)

Learn what context engineering is in AI development, why huge context windows still fail, and how to organize files for accurate outputs.

[Read the guide](https://arara.code.app.br/en/articles/context-engineering-for-ai)

### [Spec-driven development (SDD) with AI: guide](https://arara.code.app.br/en/articles/spec-driven-development-with-ai)

Understand how spec-driven development replaces ad-hoc prompting with clear contracts before generating code with AI.

[Read the guide](https://arara.code.app.br/en/articles/spec-driven-development-with-ai)

## Make room for your next idea.

Get to know Arara Code in your own project. Start with the installer and review the changes your assistant makes.

Download Arara Code
