skill-creator

Create and iterate Claude skills with SKILL.md and evaluation workflows.

Updated Feb 16, 2026
One-click install
npx skills add https://github.com/I-Synergy/AI --skill skill-creator-i-synergy
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/I-Synergy/AI/tree/main/.github/skills/skill-creator
Command: npx skills add https://github.com/I-Synergy/AI --skill skill-creator-i-synergy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This guide helps teams create new Claude Code skills, refine existing ones, and establish reliable evaluation loops to improve triggering accuracy and quality.

Core Features & Use Cases

  • Draft new skills by capturing intent, writing SKILL.md frontmatter, and outlining the instruction body.
  • Improve existing skills by editing SKILL.md, adding tests, and iterating based on feedback.
  • Set up qualitative and (when available) automated eval workflows to measure performance and safety.

Quick Start

Draft a new skill by outlining its intent and creating a SKILL.md with frontmatter and a concise instruction body that can be loaded into Claude Code.

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 iterate Claude skills from scratch?

To create Claude skills, you capture intent, draft a SKILL.md with frontmatter, outline the instruction body, and establish evaluation workflows to measure triggering accuracy and quality.

What is the best way to set up evaluation workflows for Claude skills?

Setting up evaluation workflows involves preparing test prompts to measure triggering accuracy and quality, enabling structured testing and feedback loops to refine instructions.

How does writing frontmatter improve Claude skill triggering accuracy?

Writing frontmatter defines the skill metadata, which Claude Code uses to match user intent accurately. Proper frontmatter ensures the skill triggers correctly during testing.

Can I use automated tests to measure Claude skill performance and safety?

Yes, you can establish qualitative and automated evaluation workflows to measure Claude skill performance and safety, provided you have access to evaluation guidelines and optional testing scripts.

Do I need specific scripts to run structured testing for Claude skills?

Structured testing for Claude skills requires access to evaluation guidelines and optional scripts or references to enable reliable feedback loops and measure triggering accuracy.

When should I iterate on an existing Claude skill instruction body?

You should iterate on an existing Claude skill instruction body when evaluation feedback indicates poor triggering accuracy or low quality, editing the SKILL.md and adding tests to refine performance.