agentleFS
Sign inStart free
agentcrew-academyagentcrew-academy/harness-starter-kit/AGENTS.md

harness-starter-kit

AGENTS.md41 starsChanged 18 days agoMITView on GitHub

Read CLAUDE.md in this same folder first. Everything in it about talking to someone who can't code, the questions to ask before installing, and the restart reminder applies here too. This file only covers where Codex differs. Codex has no plugin marketplace for this kit, so it always uses the by-hand route. Every hook has a Codex build under hooks/<name>/codex/, with the same logic as the Claude Code build and the interface already adapted. After installing, trigger each hook once…

  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repository6 other files it gives its agents.

The file

We split this AGENTS.md into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read CLAUDE.md in this same folder first.

    line 4
  2. Very specific:

    Every hook has a Codex build under hooks/<name>/codex/, with the same logic as the Claude Code build and the interface already adapted.

    line 10

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md
<!-- gate:allow install guide read once by the installing AI, not an always-on rules file -->

Notes for Codex

Read CLAUDE.md in this same folder first. Everything in it about talking to someone who can't code, the questions to ask before installing, and the restart reminder applies here too.

This file only covers where Codex differs. Codex has no plugin marketplace for this kit, so it always uses the by-hand route.

hooks

Every hook has a Codex build under hooks/<name>/codex/, with the same logic as the Claude Code build and the interface already adapted.

  1. Copy the scripts from hooks/<name>/codex/ into ~/.codex/hooks/, then chmod +x.
  2. Merge the blocks you need from codex-hooks-example.json into ~/.codex/hooks.json. If an event name already exists, add to its hooks array.
  3. Confirm ~/.codex/config.toml has hooks = true.
  4. Codex asks the user to trust each hook the first time it runs.
  5. Restart the session.

After installing, trigger each hook once and confirm it fires. No error is not proof it works.

Two Codex-specific notes:

  • write-manifest on Codex records which paths were touched, not the edit content, so there is no Codex version of stage-own-edits.sh. See the asymmetry note in hooks/README.md.
  • block-outgoing-comms matches tool names containing send, reply, forward, broadcast, publish, schedule or push. Codex has no "ask" decision, so every send attempt is blocked, retries included: the model shows the draft and the user sends it themselves. Tell the user this before installing it.

The interface differences between the two platforms are in the table in hooks/README.md.

skills

The six folders under skills/ are plain-text instructions and work in Codex as-is. Copy them into ~/.codex/skills/ (the .agents/skills/ links in this repo point at the same folders).

checkpoint mentions a "subagent"; use whatever Codex calls its equivalent.

Don't pretend it's installed

Say clearly what couldn't be installed and what you rewrote, and only say something works once you've tested it.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

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.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free