AGENTS.md examples from real projects

How real projects brief Codex, Cursor and every other agent that reads AGENTS.md.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Worked for most · soon

AnastasiyaW Hiding risky files

AnastasiyaWAGENTS.md

codex-claude-code-config

AnastasiyaW/codex-claude-code-config/AGENTS.md

This is a configuration system repository for AI coding agents, not an application. It collects battle-tested architectural principles, security hardening, and decision frameworks that any coding agent can drop into any project. When the user asks you to "set up this project" or "apply these principles": Every clone/copy/move/sync command requires a .claude/transfers/<id>.json contract and a # transfer-contract: marker. The contract records source, destination, operation/settings, purpose, motivation, deadline, verification, source cleanup, and the next action. The transfer Stop hook blocks unfinished…

2132d agoReviewMount

AnastasiyaWAGENTS.md

knowledge-space

AnastasiyaW/knowledge-space/AGENTS.md

This file is for AI agents that clone this repository. Read this before making any changes. A curated technical knowledge base (1343+ articles, 28 domains) designed for AI agent consumption.…

304d agoReviewMount

AnastasiyaWAGENTS.md

crossweft

AnastasiyaW/crossweft/AGENTS.md

Python 3.9+, standard library only. Before a change is done, run the checks in CONTRIBUTING.md. This repository guards its own seams with crossweft, through the same harness it installs elsewhere: This repository keeps a map of its components and of every place where two of them must agree -- routes, headers, protocol versions, DTO fields, env vars, pipe names, duplicated algorithms -- in seams/model, and crossweft check verifies the map against the code. Agents write and maintain the map, and…

421d agoReviewMount

AnastasiyaWAGENTS.md

crossweft

AnastasiyaW/crossweft/docs/agents.md

Three moments matter, whatever the agent: The stop hook checks only what changed since the last commit at which the map passed in this work tree (check --incremental), so a stop on a large repository costs the changed guards, not the whole map; the first stop, an upgrade of the tool, or a change to the map, the config or a joint plugin runs the full check. Measured on a 1,140-file map with 13 joint kinds: full check 385 s,…

421d agoReviewMount

CLAUDE.md vs AGENTS.md

About AGENTS.md

What is AGENTS.md?

An open format for instructions to coding agents, read by Codex, Cursor and others. Think of it as a README written for agents.

Where does it go?

At the repository root, with more specific files in subdirectories. Agents read the one closest to the file they're editing.

What should it contain?

Setup and test commands, code style, and the rules a new contributor would need to know.

Does Claude Code read it?

Claude Code reads CLAUDE.md. A one-line CLAUDE.md that points at AGENTS.md covers both.