Create Skill
Scaffolds high quality Claude Code skills and slash commands with correct frontmatter, conventions and worked examples.
Overview
Create Skill is a meta-skill: it helps you build other Claude Code skills and slash commands that follow established conventions: the right frontmatter, clear invocation, and reference/examples files.
How it works
It first classifies what kind of skill you want, then generates a well-structured
SKILL.md plus supporting files.
- Clarify type: Task, Research, Knowledge or Dynamic (live shell context).
- Write frontmatter: name, description, argument-hint, permissions.
- Structure the body: steps, rules, and worked examples.
- Add references:
reference.mdandexamples.mdfor depth.
Examples
| Type | Purpose | Example |
|---|---|---|
| Task | Actions with side effects | deploy, commit, publish |
| Research | Gather & synthesize | audit, deep-research |
| Knowledge | Reference context | api-conventions, style-guide |
| Dynamic | Inject live context | pr-summary, env-check |
Perfect for teams standardizing how they package prompts and workflows into reusable skills.
Installation
You need Claude Code and Git.
- Clone the skills monorepo:
git clone https://github.com/lorenzomeolav32-png/ai-assets-directory-skills - Copy the skill folder into your global skills directory:
cp -r ai-assets-directory-skills/skills/create-skill ~/.claude/skills/create-skill - Restart Claude Code, or reload the project, so it picks up the new skill.
- Ask it to scaffold a new skill to confirm it walks you through the Task/Research/Knowledge/Dynamic classification.
Related assets
Frontend Design
Generates distinctive, production-grade UI that avoids generic "AI slop," with bold typography, cohesive color, and precise motion, implemented as real code.
Graphify
Turns a codebase, plus its docs, PDFs, images, and video, into a queryable knowledge graph that a coding agent can traverse instead of grepping raw files, and never guesses silently: every edge is tagged EXTRACTED, INFERRED, or AMBIGUOUS so you know what was found versus inferred.
Caveman
Makes your coding agent answer in a stripped down, high signal style that keeps every technical fact and cuts filler, saving output tokens on chat style questions.