agentleFS
Sign inStart free
agentjidoagentjido/skills/AGENTS.md

skills

AGENTS.md16 starsChanged 6 days agoApache-2.0View on GitHub

- Use ASD-STE100 Simplified Technical English. - Use kata-ex-<task> for Elixir skills and kata-<task> for language-agnostic skills. - jido is the approved bare dispatcher. Use jido-sdk when a name refers specifically to the Jido Elixir SDK. - A small conversational or writing command can keep a distinctive bare name when the user explicitly approves the exception and the name has low collision risk. Tutor can be invoked directly or selected for clear learning intent. It must not take over ordinary…

  • 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 31,439 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 repository3 other files it gives its agents.

The file

Show its rules and text

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

14 rules. About 0 are common, 7 are distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 7

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use ASD-STE100 Simplified Technical English.

    line 3
  2. Very specific:

    Use kata-ex-<task> for Elixir skills and kata-<task> for language-agnostic skills.

    line 4
  3. Distinctive:

    Keep the plugin small. Add a skill when an actual task shows a need.

    line 9
  4. Very specific:

    Keep each skill self-contained in skills/<name>/SKILL.md.

    line 10
  5. Very specific:

    Keep author attribution in README.md, not in SKILL.md files.

    line 11
  6. Very specific:

    Use standard name and description frontmatter.

    line 12
  7. Distinctive:

    Use host-neutral instructions.

    line 13
  8. Distinctive:

    Keep the name, version, and description consistent across plugin manifests.

    line 17
  9. Distinctive:

    Do not add install tests, release automation, or distribution tooling unless requested.

    line 18
  10. Very specific:

    Treat dojo/ as preserved, opt-in evaluation tooling.

    line 19
  11. Distinctive:

    Never start a live Dojo model evaluation unless the user explicitly requests the live run and supplies or approves its profile and token or …

    line 23
  12. Very specific:

    Do not use Dojo as the default verification path for skill or documentation changes.

    line 25
  13. Very specific:

    Keep this repository public-only.

    line 27
  14. Distinctive:

    Keep marketing material outside this repository.

    line 30

≈ “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 Kata

  • Use ASD-STE100 Simplified Technical English.
  • Use kata-ex-<task> for Elixir skills and kata-<task> for language-agnostic skills.
  • jido is the approved bare dispatcher. Use jido-sdk when a name refers specifically to the Jido Elixir SDK.
  • A small conversational or writing command can keep a distinctive bare name when the user explicitly approves the exception and the name has low collision risk. Tutor can be invoked directly or selected for clear learning intent. It must not take over ordinary task execution.
  • Mark Elixir skills with metadata: {language: elixir}. No language marker means language-agnostic.
  • Keep the plugin small. Add a skill when an actual task shows a need.
  • Keep each skill self-contained in skills/<name>/SKILL.md.
  • Keep author attribution in README.md, not in SKILL.md files. Preserve required license files.
  • Use standard name and description frontmatter. Match the name to its directory.
  • Use host-neutral instructions. Do not require a specific tool API or another installed skill. An approved dispatcher can route to sibling packages bundled in the same plugin when it verifies each target and has a self-contained fallback.
  • Keep the name, version, and description consistent across plugin manifests.
  • Do not add install tests, release automation, or distribution tooling unless requested.
  • Treat dojo/ as preserved, opt-in evaluation tooling. Do not run commands from dojo/, use its catalog, run its tests, inspect its saved evaluations, or change its code unless the user explicitly requests Dojo in the current task. A skill change does not imply Dojo use.
  • Never start a live Dojo model evaluation unless the user explicitly requests the live run and supplies or approves its profile and token or call budget.
  • Do not use Dojo as the default verification path for skill or documentation changes. Use direct file inspection and checks that do not enter dojo/.
  • Keep this repository public-only. Private skills stay in a separate repository. Do not add them, or links to them, under skills/. Public files can describe host-exposed local skills but must not name a private skill.
  • Keep marketing material outside this repository.
  • Preserve the user's existing files and work when changing a target project.

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