agentleFS
arthuroc21arthuroc21/claude-skills/CLAUDE.md

claude-skills

CLAUDE.md0 starsChanged 5 months agoMITView on GitHub

claude-skills is a workspace and public GitHub repo for creating, editing, and maintaining Claude Agent Skills. It currently holds two complementary skills — kickoff (day-zero project setup) and handoff (living HANDOFF.md maintenance) — and is the home for any further skills built here. Skills are authored as plain-Markdown SKILL.md files (plus optional assets/) and target all four Claude surfaces: Chat, Cowork, Code, and Design. The SKILL.md sources are the deliverable; lightweight scripts package them for distribution without changing them —…

  • Reads credentials
  • In this repository2 other files it gives its agents.

The file

Files
CLAUDE.md

claude-skills

Project overview

claude-skills is a workspace and public GitHub repo for creating, editing, and maintaining Claude Agent Skills. It currently holds two complementary skills — kickoff (day-zero project setup) and handoff (living HANDOFF.md maintenance) — and is the home for any further skills built here. Skills are authored as plain-Markdown SKILL.md files (plus optional assets/) and target all four Claude surfaces: Chat, Cowork, Code, and Design. The SKILL.md sources are the deliverable; lightweight scripts package them for distribution without changing them — the repo doubles as a Claude Code plugin marketplace (.claude-plugin/marketplace.json + a per-skill plugin.json), and each skill is also built into its own Cowork plugin.

Working conventions

This project uses GitHub (claude-skills) for version control. Changes are committed and pushed after every significant change — a skill created or edited, a decision made or reversed, a file added or removed, a fix landed — rather than batched up. Whenever a skill changes here, every copy is kept in sync: it is committed and pushed to the claude-skills GitHub repo, the local Claude Code install at ~/.claude/skills/<skill> is refreshed, the in-repo Claude Code plugin-marketplace files (.claude-plugin/marketplace.json + each <skill>/.claude-plugin/plugin.json) are regenerated via scripts/build_marketplace.py, and one Cowork plugin per skill is rebuilt into the sibling Skills Library/ folder (<skill>.plugin, via scripts/build_cowork_plugin.py) — so no copy drifts. Commit messages are clear and specific. Secrets, API keys, and .env files are never committed; .gitignore keeps them — and packaged *.skill/*.zip/*.plugin build outputs — out of the repo. This repo is intentionally public.

Project state lives in HANDOFF.md at the repo root, maintained per the handoff skill: it is read at the start of a session and after any context compaction, and updated on every significant change, so the project can always be resumed from that one file. The HANDOFF.md format is owned solely by the handoff skill — kickoff and everything else defer to it (single source of truth).

Stack and commands

Markdown-based Agent Skills; no language toolchain, build step, or tests. A skill is a folder containing a SKILL.md (with YAML frontmatter: name, description) and optional assets/.

  • Install / sync (Claude Code): cp -r <skill> ~/.claude/skills/ (global) or <repo>/.claude/skills/ (per-project). Re-copy after every edit so the local install matches the repo.
  • Package (claude.ai upload): cd <skill> && zip -r ../<skill>.zip . && cd ..
  • Build marketplace files (Claude Code): python scripts/build_marketplace.py → regenerates .claude-plugin/marketplace.json and each <skill>/.claude-plugin/plugin.json from the skills, so the repo installs as a marketplace: /plugin marketplace add arthuroc21/claude-skills → /plugin install <skill>@claude-skills. These are tracked files, so run it before committing when a skill is added/removed or its description changes — deliberately not in the post-commit hook, which would dirty the tree. Each skill is a skills-only plugin whose SKILL.md sits at the plugin root; version is omitted so installs track the commit SHA and auto-update.
  • Build Cowork plugins: python scripts/build_cowork_plugin.py → packages each skill into its own <skill>.plugin in the sibling Skills Library/ folder, for installing individually in Cowork (each shows as <skill>:<skill>). New skills are auto-included; deleted ones have their stale .plugin removed. Run on every skill change. A git post-commit hook auto-rebuilds these (only — not the tracked marketplace files) on every commit — reinstall it after cloning with sh scripts/install-hooks.sh.
  • Conventions: Australian English, ISO dates (YYYY-MM-DD). Skills do not sync across surfaces — install separately per surface.

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.