LIVE
Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|Google's AI gives evasive answers about who won the 2020 US election28/10/26 · Google|NVIDIA introduces Nemotron 3.5 Transcribe, a promptable multilingual transcription model28/10/26 · NVIDIA|OpenAI and ServiceNow deepen partnership on enterprise workflows28/10/26 · OpenAI|Failed startups' data finds a second life training AI models28/10/26|Anthropic expands AI for Science program to 20 foundations and research institutions28/10/26 · Anthropic|Anthropic shares early results on using Claude to accelerate drug development27/10/26 · Anthropic|GEPA-Curator: Filtering Failures to Better Train LLM Agents27/10/26|Google Introduces "Gemini 4 Carbon" as Next Step in Its Model Lineup28/10/26 · Google|DeepMind unveils Genie 4, a new generation of world model27/10/26 · Google DeepMind|Anthropic files IPO prospectus, with risk disclosures taking center stage26/10/26 · Anthropic|Robot policy trained only in simulation matches top real-world systems26/10/26|OpenAI unveils GPT-6.1 Terra, a model built for computer use26/10/26 · OpenAI|Google's AI gives evasive answers about who won the 2020 US election28/10/26 · Google|NVIDIA introduces Nemotron 3.5 Transcribe, a promptable multilingual transcription model28/10/26 · NVIDIA|OpenAI and ServiceNow deepen partnership on enterprise workflows28/10/26 · OpenAI|Failed startups' data finds a second life training AI models28/10/26|Anthropic expands AI for Science program to 20 foundations and research institutions28/10/26 · Anthropic|Anthropic shares early results on using Claude to accelerate drug development27/10/26 · Anthropic|GEPA-Curator: Filtering Failures to Better Train LLM Agents27/10/26|
Skillofficial publisher

AGENTS.md (convention)

Skill : A reusable method: instructions, procedures, sometimes scripts or templates · requires any-agent

A file of instructions at the root of the repository, read by most coding agents.

The most portable contract between agents: one text file, nothing to install.

Official site
Overall score
7.9 /10
Security: nothing to flag
Utility
7.0
Reliability
6.0
Security
No caveat on record
Ease
5 min · Beginner

Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.

What it does, concretely

When multiple coding agents work on the same repository, each one may interpret project conventions, build commands, or style rules differently. AGENTS.md solves this by providing a single, predictable location to centralise all instructions intended for agents.

In practice, it is a Markdown file placed at the root of the repository. The agent reads it at the start of each task, extracts build commands, style rules, testing instructions, and any other relevant context, then adapts its behaviour accordingly. In a monorepo, nested AGENTS.md files can be placed inside each subfolder: the agent always reads the file closest to the directory it is working in.

This format is designed for developers who use coding agents such as Codex, Cursor, Gemini CLI, GitHub Copilot, or any of the other compatible tools. It suits both individual projects and large shared codebases.

On the other hand, AGENTS.md offers little value to those who do not use coding agents in their workflow, or to those working on projects without a structured Git repository.

What it offers

  • ·A single Markdown file at the repository root centralises all instructions for coding agents.
  • ·Compatible with more than 20 coding agents, including Codex, Cursor, Gemini CLI, GitHub Copilot, Aider, Devin, Windsurf, and Jules.
  • ·Supports nested AGENTS.md files in monorepo subfolders, with the closest file taking precedence.
  • ·Test commands listed in the file are executed automatically by the agent, which fixes errors before completing the task.
  • ·Free Markdown format with no required fields: headings, sections, and content are fully customisable.
  • ·When multiple files conflict, the one closest to the relevant directory wins; instructions typed directly in the agent chat override everything.
  • ·Open format, now stewarded by the Agentic AI Foundation under the Linux Foundation.

In practice

Price
Free
Commercial use
Commercial use: yes
Available on
codex, cursor, gemini-cli, copilot
Host
any-agent
FR interface
No

Security and data

Rating
Security: nothing to flag
Your data
Data stays local
Maintenance
Active

Links and repository checked on 25 October 2026 · added on 1 August 2025

How to install AGENTS.md (convention)

1. Create a file named AGENTS.md at the root of your Git repository. 2. Add relevant sections: project overview, build and test commands, style rules, commit conventions. 3. For a monorepo, add additional AGENTS.md files inside each relevant subfolder. 4. If you use Aider, add 'read: AGENTS.md' to your .aider.conf.yml file. 5. If you use Gemini CLI, add to .gemini/settings.json: { "context": { "fileName": "AGENTS.md" } }. 6. Commit the file to your repository: it contains only text, and no sensitive data should be included.

In the same category

Whole category Frame (methods, guardrails) →