• Picks
  • Discover
  • Alternatives
  • About

© 2026 Willet

  • Newsletter
  • About
  • FAQ
  • Terms
  • Privacy & Cookies
Join the Discord (opens in a new tab)
AGENTS.md avatar

AGENTS.md

Open file convention for giving coding agents project instructions

  • Agent skills and plugins
  • AI coding agents
View on GitHub
Website

What it is

AGENTS.md is a plain Markdown file format placed in a repository to tell AI coding agents how to set up, build, test, and submit changes to a project. It works like a README written for agents, covering things such as dev environment commands, test steps, and pull request rules. Maintainers add one so that different coding agents pick up the same project conventions. The repo also holds the Next.js site that explains the format with examples.

Images

Facts

Stars
25,014 GitHub stars.
25,014

Forks
1,915
Commits
38
Contributors
21
Language
TypeScript
License
MIT
Last commit
Sep 10, 2026
On GitHub since
August 2025
Self-hostable
No
Tags
  • coding-agent
  • markdown
  • documentation
  • context-engineering

Last synced from GitHub on Oct 20, 2026

The 10 best overlooked open-source projects, every Monday.

Share and more

Comments

No comments yet. Start the conversation.

Similar projects

  • Agent Skills avatar

    Agent Skills by agentskills

    Open specification for packaging skills that AI agents load on demand

    26.3k
  • Ahmad Othman Ammar Adi avatar

    Planning with Files

    File-based planning skill that keeps coding agents on track across resets

    27.6k
  • Microsoft avatar

    HVE Core

    Agents, instructions, and skills that structure GitHub Copilot work

    1.5k

What people are saying

Updated Oct 21, 2026
  • Hacker NewsAGENTS.md – Open format for guiding coding agentsAug 20, 2025 · 837 points · 382 comments (opens Hacker News in a new tab)
  • Bluesky“echo "if I see a sha256 again you're going to robot jail" >> AGENTS.md”Oct 20, 2026 · 6 likes · 1 reply (opens Bluesky in a new tab)
  • Bluesky“Codex has a default 32 KiB limit for AGENTS.md files combined. So some may never reach the agent if the file is too big. I cut out anything that's obvious…”Oct 19, 2026 · 2 likes · 5 replies (opens Bluesky in a new tab)
  • Bluesky“skimming some of the recent transcripts, the bigger issue with my system is that some memory files are too general and too big; a few of them should probably…”Oct 19, 2026 · 2 likes · 1 reply (opens Bluesky in a new tab)
  • Bluesky“My custom AGENTS.md, rules and skills go beyond naming conventions. They guide architectural decisions, reduce duplication, challenge bad abstractions…”Oct 20, 2026 · 1 like · 1 reply (opens Bluesky in a new tab)
  • Bluesky“I feel like one could set traps inside the files they use, namely CLAUDE.md and AGENTS.md Probably won't stop them for long, I'm afraid (maybe make the…”Oct 19, 2026 · 1 like · 1 reply (opens Bluesky in a new tab)
  • Bluesky“AGENTS.md gives very generic guidance and shows where to find more info. Something like "project is divided into folders with each corresponding to one…”Oct 19, 2026 · 1 like · 1 reply (opens Bluesky in a new tab)
  • Bluesky“This is a first priority rule I have in every AGENTS.md or related doc. It’s maddening.”Oct 18, 2026 · 1 like · 1 reply (opens Bluesky in a new tab)
  • Bluesky“As indicated in the post, an HN thread asked whether user-level files count, and the Claude Code memory docs answer it. To check, run /memory: since v2.1.280…”Oct 19, 2026 · 0 likes · 1 reply (opens Bluesky in a new tab)
  • Bluesky“AGENTS.md per stack makes sense since the test command alone differs so much: cargo test, pytest, go test ./... An agent guessing that wrong burns a whole…”Oct 19, 2026 · 0 likes · 1 reply (opens Bluesky in a new tab)
  • Bluesky“OpenAI says trim bloated Skills, AGENTS.md, and prompts for GPT-6, define clear done criteria, and authorize local tests so agents work autonomously. How will…”Oct 19, 2026 · 1 like · 0 replies (opens Bluesky in a new tab)
  • Bluesky“RT @marcos_placona: So I’ve been using @t3dotcodes for a few weeks now and love using projects. I couldn’t find a way to set initial context per project, so I…”Oct 19, 2026 · 0 likes · 1 reply (opens Bluesky in a new tab)
  • Bluesky“You could Google it. Does your AGENTS.md file allow you to do that? What they say about em dashes is that, while they are used by real human beings in…”Oct 17, 2026 · 0 likes · 1 reply (opens Bluesky in a new tab)
  • Bluesky“very token efficient ways to catch issues ahead of time. 3. prose: comments, rules, AGENTS.md, skills - anything that can't be enforced and requires agent…”Oct 16, 2026 · 0 likes · 1 reply (opens Bluesky in a new tab)
  • Bluesky“Would you believe I had lines to that effect already in my AGENTS.md and an alternative non-Docker make command I want it to use instead?”Oct 16, 2026 · 0 likes · 1 reply (opens Bluesky in a new tab)
  • Bluesky“Two details from codex-rs/core/src/agents_md.rs: the 32 KiB is spent root-first, so the file that gets cut is the deepest, most folder-specific AGENTS.md. And…”Oct 20, 2026 · 0 likes · 0 replies (opens Bluesky in a new tab)
  • Bluesky“Cargo, because the safety rules are the part of a stack's AGENTS.md that isn't just a command list. Build scripts and proc macros mean the agent can run…”Oct 20, 2026 · 0 likes · 0 replies (opens Bluesky in a new tab)
  • Bluesky“New in our free AGENTS.md generator: paste your package.json, pyproject.toml, Cargo.toml or go.mod and it fills in your real install, test, lint and build…”Oct 19, 2026 · 0 likes · 0 replies (opens Bluesky in a new tab)
  • Bluesky“Yeah, specify where you want temp files to go in your AGENTS.md for sure! Have a practice! Have it git ignored!”Oct 19, 2026 · 0 likes · 0 replies (opens Bluesky in a new tab)
  • Bluesky“Instruction files are executable config in practice. I'd put CODEOWNERS on AGENTS.md and the skills directory and review changes to them like CI config, since…”Oct 19, 2026 · 0 likes · 0 replies (opens Bluesky in a new tab)