agentleFS
Sign inStart free
ac5tinac5tin/skills/AGENTS.md

skills

AGENTS.md0 starsChanged 35 days agoMITView on GitHub

A collection of reusable agent skills (Markdown-based instruction sets) for AI coding tools. No build system or dependencies. After changing packaging, run claude plugin validate . (a missing-version warning is expected).

  • 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 repository9 other files it gives its agents.

The file

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

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    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
  1. Very specific:

    No tags or version bumps. The only release automation is .github/workflows/pin-plugin-sha.yml, which re-pins source.sha when plugin content …

    line 29
  2. Very specific:

    No build system or dependencies.

    line 33
  3. Very specific:

    Keep skills focused and concise — one concern per skill.

    line 38
  4. Very specific:

    Use descriptive kebab-case directory names for new skills (e.g., my-new-skill/).

    line 39
  5. Distinctive:

    Follow existing formatting: H2 sections for top-level topics, H4 for subsections, fenced code blocks for examples.

    line 40
  6. Distinctive:

    Do not duplicate guidance already covered by an existing skill.

    line 41

≈ “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

Agent Skills Repository

A collection of reusable agent skills (Markdown-based instruction sets) for AI coding tools.

Structure

  • skills/ — Each subdirectory is a skill containing a SKILL.md with YAML frontmatter (name, description) and instructions.
    • implementation-plan/ — Required plan structure and the definition-of-done mapping.
    • plan-drive/ — Ordered todo list from an approved plan, with the wait rule and the no-idle rule.
    • plan-verify/ — Executable verification scripts and run order from an approved plan.
    • code-quality/ — General clean-code best practices (KISS, DRY, simplicity).
    • golang-code/ — Go coding standards: error handling, naming, anti-patterns, pointers, function signatures, linting.
    • go-service-scaffold/ — Scaffold a Go microservice (includes reference templates).
    • next-steps/ — End finished tasks with clear next actions and follow-ups.
    • orchestrate/ — Multi-agent orchestration: subagent fan-out, isolated parallel reviews, arbitration. Triggered only via explicit invocation.
    • trivy-scanner/ — Run Trivy vulnerability scans via Podman.
  • agents/ — Agent templates that are not skills. dr-autobot/ creates autonomous PIC driver bots; its files are read on demand and never stored in memory.
  • docs/ — Human-facing guides. dr-autobot.md covers install, first run and the driver contract.
  • Plugin packaging (no skills/ changes needed; new skills are picked up automatically):
    • .claude-plugin/ — Claude Code, Claude Desktop, ZCode. version is deliberately omitted; the plugin entry pins a source.sha (github object) that CI bumps per release — the pinned SHA is the version and drives ZCode update detection.
    • .codex-plugin/, .agents/plugins/ — Codex. version is a constant 0.0.0 because Codex requires the field.
    • .grok-plugin/ — Grok Build. marketplace.json is a self-hosted catalog whose single url source points back at this repo with no pinned SHA, so it tracks main; plugin.json is the plugin manifest (skills points at ./skills/). No plugin-index.json — that is generated only inside the official xai-org/plugin-marketplace repo.
    • .cursor-plugin/ — Cursor and Grok Bot share this format. marketplace.json is a single-plugin catalog with source: "."; its schema allows only name, source, description, minClientVersions per entry. plugin.json author takes name/email only, so no url. No version, consistent with the no-version convention.
    • package.json (pi.skills) — Pi.
    • .opencode/plugins/ac5tin-skills.js + root index.js — OpenCode v2; registers every skills/*/SKILL.md.
    • No tags or version bumps. The only release automation is .github/workflows/pin-plugin-sha.yml, which re-pins source.sha when plugin content changes; otherwise a new commit is a new release.

Build / Test

No build system or dependencies. After changing packaging, run claude plugin validate . (a missing-version warning is expected).

Conventions

  • Skills are plain Markdown files named SKILL.md with YAML frontmatter (name, description).
  • Keep skills focused and concise — one concern per skill.
  • Use descriptive kebab-case directory names for new skills (e.g., my-new-skill/).
  • Follow existing formatting: H2 sections for top-level topics, H4 for subsections, fenced code blocks for examples.
  • Do not duplicate guidance already covered by an existing skill; extend or reference it instead.

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