Can I use Skill Creator to benchmark existing skills?

Yes, it supports taking snapshots of existing skills and running side-by-side evaluations against new versions to measure quantitative and qualitative improvements.

How are skill evaluations performed?

The tool spawns subagents to run tasks both with and without the skill, saving outputs to a dedicated workspace for comparison and variance analysis.

What is a Claude Code Skill?

A Claude Code Skill is a specialized capability that provides domain-specific guidance, best practices, and implementation patterns to enhance Claude's performance on specific tasks.

How does Skill Creator help with skill triggering?

It includes strategies to optimize skill descriptions, making them 'pushy' and context-aware to ensure Claude activates the skill precisely when it is most useful.

Do I need to be an expert coder to create skills?

No, Skill Creator is designed to guide users with varying levels of technical expertise, providing brief explanations of technical terms like JSON or assertions when necessary.

Skill Creator

Name: Skill Creator
Author: hg-pyun

byhg-pyun

•

開発者ツール

Automates the lifecycle of creating, testing, and optimizing specialized Claude Code skills.

Skill Creator provides a robust framework for developing high-performance Claude Code skills through a structured, iterative workflow. It guides users from initial intent capture and research through drafting concise SKILL.md instructions to advanced performance benchmarking. By spawning subagents for side-by-side baseline comparisons and drafting quantitative assertions, it ensures that every skill is not only functional but optimized for accuracy, reliability, and efficient triggering.

主な機能

01Standardized workspace management for version-controlled iterations

021 GitHub stars

03Quantitative assertion generation for objective skill evaluation

04Iterative skill development from drafting to large-scale testing

05Triggering optimization to prevent skill 'undertriggering'

06Automated side-by-side performance benchmarking against baselines

ユースケース

01Benchmarking and improving the accuracy of existing Claude skills

02Optimizing skill metadata to ensure reliable activation during conversations

03Building custom domain-specific skills for proprietary coding workflows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add hg-pyun/claude-code-marketplace skill-creator

For use in Claude.ai and ChatGPT

主な機能

01Standardized workspace management for version-controlled iterations

021 GitHub stars

03Quantitative assertion generation for objective skill evaluation

04Iterative skill development from drafting to large-scale testing

05Triggering optimization to prevent skill 'undertriggering'

06Automated side-by-side performance benchmarking against baselines

ユースケース

01Benchmarking and improving the accuracy of existing Claude skills

02Optimizing skill metadata to ensure reliable activation during conversations

03Building custom domain-specific skills for proprietary coding workflows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add hg-pyun/claude-code-marketplace skill-creator

For use in Claude.ai and ChatGPT