
Skill Developer Guide
Teaches the agent how to create, configure and debug Claude Code skills and their hooks
- Stars
- 10,043
- Licence
- MIT
- Files
- 7 files
What it does
Once installed, the agent knows how to build and maintain its own skills for Claude Code. It explains the SKILL.md format, the skill-rules.json configuration, the trigger types (keywords, intent patterns, file paths, content patterns), enforcement levels (block, suggest, warn), and the hook scripts that decide when skills appear or block edits. It also gives step-by-step testing commands and the 500-line guidance for keeping skills lean.
Great for
- Developers setting up the auto-activation skill system in their own Claude Code project
- Teams writing custom skills who want consistent structure and trigger tuning
- Anyone debugging why a skill did or did not activate
Skip it if
- Non-developers looking for everyday help: it is about building skills, not doing a task, and it assumes familiarity with hooks, JSON and regex
- Projects that copy the hook scripts without reviewing them: the described setup runs TypeScript hooks that can block file edits and read session data, so unreviewed installation can change how the agent behaves
Try asking
- “Create a new skill for our database schema conventions and wire it into skill-rules.json”
- “Why isn't my skill activating when I mention deployment? Help me debug the trigger patterns”
- “Explain the difference between block and suggest enforcement and which to use for a migration guardrail”
- “Refine the intent patterns in skill-rules.json to reduce false positives”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 7 original files, the MIT licence and an attribution note.
skill-developer.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip skill-developer.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip skill-developer.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Use this skill only in a project where you intend to maintain the hook setup, and review the .claude/hooks scripts before enabling them
- Test new triggers with the provided echo/npx commands before relying on them, and keep block enforcement for a small number of critical guardrails
- Keep SKILL.md under 500 lines and move detail into reference files, as the guide itself recommends
What's inside
- ADVANCED.md
- HOOK_MECHANISMS.md
- PATTERNS_LIBRARY.md
- SKILL_RULES_REFERENCE.md
- SKILL.md
- TRIGGER_TYPES.md
- TROUBLESHOOTING.md
Redistributed unmodified from diet103/claude-code-infrastructure-showcase under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.