Skip to content

AGENTS.md

Latest commit

 

History

History
35 lines (28 loc) · 1.52 KB

AGENTS.md

File metadata and controls

35 lines (28 loc) · 1.52 KB
Raw

AGENTS.md

Guidelines for any coding agent working in this repository. This is the shared source of truth; CLAUDE.md imports it via @AGENTS.md.

What this repo is

A personal skills library for coding agents, published as a plugin for Claude Code, Codex, and Cursor. The same skills/ directory is shared by all three — each *-plugin/plugin.json points skills at ./skills/, so a skill is written once and works everywhere.

Rules

  • A skill is a folder under skills/<category>/<name>/ containing a SKILL.md. The frontmatter name MUST equal the folder name (kebab-case).
  • description is the trigger the agent matches on — write it as Use when ..., and be specific about when to fire AND when not to.
  • One capability per skill. Keep the body imperative and short.
  • Bump version in ALL FOUR manifests together (they must stay in sync): .claude-plugin/marketplace.json (plugin entry), .claude-plugin/plugin.json, .codex-plugin/plugin.json, .cursor-plugin/plugin.json.
  • All manifests are JSON — validate before committing (python3 -m json.tool < file).
  • Never commit secrets. This repo is public.

Layout

.claude-plugin/marketplace.json   # Claude marketplace catalog (name: pmueller-skills)
.claude-plugin/plugin.json        # Claude plugin manifest (name: skills)
.codex-plugin/plugin.json         # Codex manifest  (skills -> ./skills/)
.cursor-plugin/plugin.json        # Cursor manifest (skills -> ./skills/)
skills/<category>/<name>/SKILL.md # skills (shared by all agents)