skill-creator

Guide Claude Code skill development through SKILL.md creation and test cases.

Updated Jun 3, 2026
One-click install
npx skills add https://github.com/gclm/skills-hub --skill skill-creator-gclm
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/gclm/skills-hub/tree/main/skill-creator
Command: npx skills add https://github.com/gclm/skills-hub --skill skill-creator-gclm

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires scripts.validate.py, scripts.aggregate_benchmark.py, scripts.generate_report.py, scripts.run_loop.py, eval-viewer.generate_review.py, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill empowers users to create, modify, and evaluate Claude Code skills, streamlining the development process and enabling efficient skill iteration.

Core Features & Use Cases

  • Skill Creation: Assist in defining, writing, and refining Claude Code skills.
  • Skill Editing & Optimization: Iterate on existing skills based on user feedback and performance metrics.
  • Skill Evaluation: Utilize quantitative and qualitative methods to assess skill effectiveness.
  • Use Case: Suppose you want to develop a skill for generating code snippets. This Skill would guide you through every step, from defining the skill's purpose to running evaluations and optimizing the skill based on feedback.

Quick Start

Start the skill creator by initiating the process to create a new skill: 'create skill new skill_name'.

Frequently Asked Questions about skill-creator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a Claude Code skill from scratch?

To create a Claude Code skill, initiate the process with the command 'create skill new skill_name' to start guided intent capture, write the SKILL.md file, and run test cases for iterative improvement.

How does iterative improvement work for Claude Code skill development?

Iterative improvement for Claude Code skill development utilizes subagents for parallel execution and benchmarking, optimizing skill triggering accuracy by incorporating user feedback and quantitative performance metrics continuously.

Can I evaluate Claude Code skills using quantitative metrics?

Yes, you can evaluate Claude Code skills using quantitative methods by running validation scripts and aggregating benchmarks to generate reports that assess skill effectiveness and triggering accuracy.

Do I need subagents to run skill evaluation benchmarks?

Yes, subagents are required for parallel execution, analysis, and benchmarking during skill evaluation, enabling the system to process multiple test cases and aggregate performance data efficiently.

What's the best way to optimize skill triggering accuracy in Claude Code?

The best way to optimize skill triggering accuracy is through iterative improvement loops that run test cases, aggregate benchmark data, and incorporate user feedback to refine the SKILL.md definition.

Why does my Claude Code skill need a SKILL.md file?

A SKILL.md file is needed to define the Claude Code skill's purpose and triggering logic, serving as the core configuration that guides the skill's behavior and enables accurate evaluation during development.