agentleFS
Sign inStart free
undeemedundeemed/manifest/AGENTS.md

manifest

AGENTS.md1 starsChanged 3 months agoMITView on GitHub

This repo IS a skill. SKILL.md at the root is the product; everything else is packaging and proof. - SKILL.md is canonical. skills/manifest/SKILL.md is a plugin-layout copy. After editing the root file: cp SKILL.md skills/manifest/SKILL.md. CI diffs them. - No em dashes anywhere. Use plain -. CI greps for it. - Every change to SKILL.md needs an eval run before merge: python3 evals/run.py --model sonnet --judge-model sonnet --workers 6 Pass criteria: vs the committed baseline, helpless(A=0) not up, fabricated(H=0) not…

  • 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 31,542 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 repository2 other files it gives its agents.

The file

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

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    After editing the root file: cp SKILL.md skills/manifest/SKILL.md.

    line 9
  2. Very specific:

    No em dashes anywhere. Use plain -.

    line 10
  3. Very specific:

    Every change to SKILL.md needs an eval run before merge: python3 evals/run.py --model sonnet --judge-model sonnet --workers 6 Pass …

    line 11

≈ “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

manifest - agent instructions

This repo IS a skill. SKILL.md at the root is the product; everything else is packaging and proof.

Working on this repo

  • SKILL.md is canonical. skills/manifest/SKILL.md is a plugin-layout copy. After editing the root file: cp SKILL.md skills/manifest/SKILL.md. CI diffs them.
  • No em dashes anywhere. Use plain -. CI greps for it.
  • Every change to SKILL.md needs an eval run before merge: python3 evals/run.py --model sonnet --judge-model sonnet --workers 6 Pass criteria: vs the committed baseline, helpless(A=0) not up, fabricated(H=0) not up, disguised_solvable A/H not down. Commit the results JSON.
  • The honesty floor section of the skill is load-bearing. Do not weaken it to boost attempt scores.
  • Eval harness is stdlib-only Python 3.9+. Keep it that way; no requirements.txt.

Layout

PathRole
SKILL.mdThe skill (canonical)
skills/manifest/SKILL.mdPlugin copy, must match root
commands/manifest.md/manifest slash command
.claude-plugin/Claude Code plugin + marketplace manifests
evals/prompts.json5 categories, 17 prompts
evals/run.pyA/B runner + LLM judge
evals/results/Committed benchmark evidence
install.shManual installer for ~/.claude

Testing your changes to the skill quickly

Smoke (4 calls): python3 evals/run.py --category truly_impossible --limit 1 --model haiku --judge-model haiku

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