agentleFS
Sign inStart free
ADWilkinsonADWilkinson/astack/AGENTS.md

astack

AGENTS.md0 starsChanged 8 days agoMITView on GitHub

This repo is the astack skill suite. These notes are for agents changing the suite itself, not for using it.

  • 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 31,843 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 repository11 other files it gives its agents.

The file

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

6 rules. About 1 is common, 1 is distinctive, and 4 are very specific to this project.

  • ≈1

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Very specific:

    Every skill folder must work on its own.

    line 6
  2. Very specific:

    When you add, rename or remove a skill, update catalog.json, .claude-plugin/plugin.json and the README table, then run bun run metadata.

    line 7
  3. Very specific:

    Keep SKILL.md short and concrete: a specific description, the useful procedure, real boundaries and what "done" looks like.

    line 8
  4. Very specific:

    Run bun run check before committing.

    line 10
  5. Common:

    Never commit secrets, private transcripts, customer data or anyone's personal account details.

    line 12
  6. Distinctive:

    Use Bun for repo tooling. No telemetry, install hooks, remote code or MCP registration in the package.

    line 13

≈ “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

Working on astack

This repo is the astack skill suite. These notes are for agents changing the suite itself, not for using it.

  • One canonical copy of each skill lives in skills/<name>/. Claude Code and Codex read the same tree.
  • Every skill folder must work on its own. Links stay inside the folder. Scripts used by more than one skill live in shared/ and are copied in by bun run metadata; edit the shared copy, never a skill's copy.
  • When you add, rename or remove a skill, update catalog.json, .claude-plugin/plugin.json and the README table, then run bun run metadata.
  • Keep SKILL.md short and concrete: a specific description, the useful procedure, real boundaries and what "done" looks like. Put conditional material in references/; another skill is not a mandatory stage.
  • Scripts are plain Node 22 with no dependencies, print usage with --help and use meaningful exit codes. bun run smoke runs every installed script's --help; prove new script behaviour with a run against a synthetic fixture or an eval, not a new unit test. tests/scripts.test.ts stays as the optional bun test suite.
  • Run bun run check before committing. It is the CI gate: structure validation, the package-boundary tests and the packed-CLI smoke (see CONTRIBUTING.md). Run slop-scan.mjs on any prose you change; the repo should pass its own deslop.
  • A change to how a skill behaves should be backed by a before/after eval (see evals/), not just a passing check. Repeat representative tasks and inspect individual regressions even when the total improves. Keep independent review for risk or judgment that deterministic checks cannot settle.
  • Never commit secrets, private transcripts, customer data or anyone's personal account details. Fixtures are synthetic.
  • Use Bun for repo tooling. No telemetry, install hooks, remote code or MCP registration in the package.

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.

Reviews can't be read right now.

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