agentleFS
Sign inStart free
ryankoleanryankolean/summit-claude-skills/AGENTS.md

summit-claude-skills

AGENTS.md3 starsChanged 7 months agoMITView on GitHub

Agent-facing guide for the Summit Claude Skills repo. Human-facing docs live in README.md. A curated directory of third-party Claude skills, MCP servers, and prompt libraries, plus a growing set of Summit-authored custom skills under skills/. The repo is mostly documentation (the README is the primary artifact) with skill definitions as standalone SKILL.md files organized one per directory. No build or test pipeline. Useful scripts: Run ad hoc from the repo root. - One skill per directory under skills/. Every skill…

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

8 rules. About 0 are common, 5 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    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. Distinctive:

    No build or test pipeline.

    line 18
  2. Distinctive:

    Run ad hoc from the repo root.

    line 24
  3. Very specific:

    External entries: must meet the bar in CONTRIBUTING.md (100+ stars, active maintenance, OSS license).

    line 46
  4. Very specific:

    Do not add emojis to the README or any SKILL.md

    line 51
  5. Distinctive:

    Do not rewrite the install-prompt block from scratch — it's numbered and ordered.

    line 52
  6. Distinctive:

    Do not create a skill directory without also updating the README table and install-prompt list

    line 53
  7. Very specific:

    Do not commit node_modules/, local caches, or editor state

    line 54
  8. Distinctive:

    Do not redistribute third-party skill source code — link to the origin repo instead

    line 55

≈ “Common” is an estimate: at least 1 in 5 of the 36,368 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

AGENTS.md — summit-claude-skills

Agent-facing guide for the Summit Claude Skills repo. Human-facing docs live in README.md.

Project overview

A curated directory of third-party Claude skills, MCP servers, and prompt libraries, plus a growing set of Summit-authored custom skills under skills/. The repo is mostly documentation (the README is the primary artifact) with skill definitions as standalone SKILL.md files organized one per directory.

Tech stack

  • Markdown (README is the directory; each custom skill is its own SKILL.md)
  • Shell scripts in scripts/ (.sh)
  • Static assets in assets/
  • Install-prompt text in install-prompts/ALL-PROMPTS.md

Commands

No build or test pipeline. Useful scripts:

  • scripts/check-broken-links.sh — validate external links in README
  • scripts/update-stars.sh — refresh GitHub star counts (if present)
  • scripts/install-local.sh <skill>... | --all — copy skills into ~/.claude/skills; re-run after pulling skill changes

Run ad hoc from the repo root.

Structure

  • README.md — top-level directory; source of truth for the curated list and custom-skills table
  • CONTRIBUTING.md — submission rules for third-party entries
  • skills/{skill-name}/SKILL.md — one directory per Summit-authored skill
  • install-prompts/ALL-PROMPTS.md — canonical install-prompt block mirrored in README
  • templates/ — reusable starter files (personal CLAUDE.md template, etc.) — not skills; copy-and-fill
  • assets/ — logos and static images
  • scripts/ — maintenance shell scripts

Conventions

  • One skill per directory under skills/. Every skill ships a SKILL.md with frontmatter (name, description, when_to_use / triggers) followed by rules and examples.
  • Two places to update when adding a Summit skill:
    1. A new row in the "Summit Custom Skills" table in README.md
    2. A numbered entry in the install-prompt block (both in README.md and install-prompts/ALL-PROMPTS.md) Renumber later entries (e.g., MENTOR SKILLS, MENTOR COUNCIL) to keep ordering consistent.
  • No emojis anywhere — README, SKILL.md, commit messages, or install prompts.
  • Commit messages: conventional style (feat:, fix:, docs:, chore:), short imperative first line.
  • Direct commits to main — no feature branches, no PRs, unless the change is risky or collaborative.
  • External entries must meet the bar in CONTRIBUTING.md (100+ stars, active maintenance, OSS license).
  • Third-party skills are never redistributed — we only link. Do not copy a skill's source into this repo.

Do not

  • Do not add emojis to the README or any SKILL.md
  • Do not rewrite the install-prompt block from scratch — it's numbered and ordered; insert/renumber in place
  • Do not create a skill directory without also updating the README table and install-prompt list
  • Do not commit node_modules/, local caches, or editor state
  • Do not redistribute third-party skill source code — link to the origin repo instead

Related repos

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