skill-creator

Create and refine Claude skills with SKILL.md drafting and iterative evals.

Updated May 22, 2025
One-click install
npx skills add https://github.com/curtjen/.gir --skill skill-creator-curtjen
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/curtjen/.gir/tree/main/skills/skill-creator
Command: npx skills add https://github.com/curtjen/.gir --skill skill-creator-curtjen

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a repeatable framework to create new Claude skills, refine existing ones, and measure performance through structured evals and benchmarks. It guides users from drafting SKILL.md to running iterations and updating triggering descriptions based on feedback.

Core Features & Use Cases

  • End-to-end skill creation: draft SKILL.md, assemble references, scripts, and assets, and define evaluation prompts.
  • Evaluation loop orchestration: run eval prompts, capture results, and generate review data to iterate quickly.
  • Trigger optimization: improve skill triggering descriptions based on feedback to increase activation accuracy.

Quick Start

Write a new SKILL.md with a clear name and description, wire up a basic evaluation prompts, and begin the iterative improvement cycle.

Frequently Asked Questions about skill-creator

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

FAQPage Schema
How do I create and evaluate a new skill from scratch?

To create a skill, draft a SKILL.md file with a clear name and description, assemble references and scripts, define evaluation prompts, and run iterative evals to measure performance and drive continual enhancement.

What is the best way to improve skill triggering descriptions?

Trigger optimization improves skill triggering descriptions by capturing evaluation results and generating review data, allowing you to iteratively update descriptions based on feedback to increase activation accuracy.

How does the iterative evaluation loop work for refining skills?

The iterative evaluation loop works by running test prompts against your skill, capturing the results, and generating structured review data to identify performance gaps and drive quick, measurable refinements.

Can I use this skill creation workflow to improve existing modules?

Yes, you can use this workflow to improve existing skill modules by running structured evals and benchmarks against current test prompts, then applying the feedback loop to evolve their performance.

What do I need to start building a skill module?

To start building a skill module, you need a drafted SKILL.md file, any necessary references, scripts, or assets, and a set of defined evaluation prompts to establish a disciplined testing process.