agentleFS
Sign inStart free
wyeworkswyeworks/accountable-review/AGENTS.md

accountable-review

AGENTS.md42 starsChanged 41 days agoMITView on GitHub

Working rules for this repository live in .claude/CLAUDE.md. Read it in full before editing anything here. It is the canonical home for every rule that spans files — what the two skills are, the invariants several files have to agree on, and the register to write in — and this file deliberately restates none of it. That is the point rather than an economy: CLAUDE.md § One canonical home is the page's rule and that file's own, and a second…

The file

Create a free account to mount
Files
AGENTS.md

AGENTS.md

Working rules for this repository live in .claude/CLAUDE.md. Read it in full before editing anything here. It is the canonical home for every rule that spans files — what the two skills are, the invariants several files have to agree on, and the register to write in — and this file deliberately restates none of it.

That is the point rather than an economy: CLAUDE.md § One canonical home is the page's rule and that file's own, and a second copy of the rules goes stale where a pointer cannot.

What is different when you work here from Codex

Three facts CLAUDE.md cannot state, because it is written for the host that has the Claude CLI.

  • bin/evals offline cannot come back fully clean without claude on PATH. Its manifest suite is claude plugin validate . --strict, gated on the binary being there, and a suite that did not run is reported as SKIP and makes the run unclean on purpose — not a pass. Everything else in the suite is shell and Ruby and runs anywhere. CI runs the whole thing on every push, so a skipped manifest locally is a deferred check rather than a missing one.
  • skills/setup-ci/ and ci/ generate and run a workflow that invokes Claude, with Anthropic credentials. Codex does not run them yet. Editing that half from Codex is fine; the deterministic tests under skills/setup-ci/tests/ cover it without a model.
  • Whole-page evals need a model, so the same split applies: bin/evals offline is the part you can always run, and bin/evals e2e generates through ci/generate-review-map.sh, which invokes Claude with an Anthropic credential, so it does not run from Codex.

The product's own Codex support — the Codex plugin and marketplace, and linking review-map into a Codex skills directory — is a different subject, and docs/codex.md owns it.

CONTRIBUTING.md is how to run things. CLAUDE.md is why the rules exist.

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