skills.rest
CZ-X-Ya
CZ-X-YaCOMMUNITY
Explore all skills by this author →
GitHub

skill-creator

Generates Claude Code skill drafts and iteratively improves them via structured evaluations and benchmarks.

Updated May 11, 2026
Search Tags:#quality assurance#skill creation#benchmarking#Claude Code#evals#trigger optimization#instruction refinement
One-click install
npx skills add https://github.com/CZ-X-Ya/oh-my-csharp-skills --skill skill-creator-cz-x-ya
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/CZ-X-Ya/oh-my-csharp-skills/tree/main/skill-creator
Command: npx skills add https://github.com/CZ-X-Ya/oh-my-csharp-skills --skill skill-creator-cz-x-ya

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and assets (resource) components.

What problem does it solve?

Skill creation and iteration is slow and guessy: you need a clear way to draft a new skill, test it against realistic prompts, measure trigger accuracy and performance, and then improve the instructions based on evidence.

Core Features & Use Cases

  • Draft from scratch or iterate on an existing skill: guides the user through producing a strong SKILL.md, including trigger description quality.
  • Run evals and compare with baselines: evaluates “with-skill” versus “without-skill” behavior across a test set.
  • Quantitative benchmarking with variance: aggregates results into pass-rate/time/token summaries for analysis.
  • Reviewer workflow for qualitative feedback: launches a viewer so you can inspect outputs and add structured feedback.
  • Description optimization: generates and tests candidate trigger descriptions to improve triggering accuracy.

Quick Start

Use skill-creator when you want to turn an idea into a new Claude Code skill or to improve an existing one by running a small eval set, reviewing results, and iterating until triggering and outcomes are reliable.

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 benchmark Claude Code skills?▼

To create and benchmark Claude Code skills, you draft a SKILL.md, build test prompts, and run structured evaluations comparing with-skill versus without-skill behavior to measure performance.

What is the best way to improve trigger accuracy for Claude skills?▼

Improving trigger accuracy requires description optimization, which generates and tests candidate trigger descriptions against a test set to tune the skill description for reliable triggering.

How do I evaluate the performance of a Claude Code skill?▼

Evaluating a Claude Code skill involves running evals, grading outputs, and aggregating results into pass-rate, time, and token summaries with variance for quantitative benchmarking analysis.

Can I iterate on an existing skill using structured evaluations?▼

Yes, you can iterate on an existing skill by running an evidence-driven workflow that executes evals, launches a viewer for qualitative feedback, and loops instruction refinement until outcomes are reliable.

What is the process for refining skill instructions based on evidence?▼

Refining skill instructions based on evidence involves executing evaluations, grading the results, reviewing outputs via a viewer, and looping iterations using provided scripts to guide improvements.

Related Skills

Explore More Software Engineering Skills→
dparedesidparedesi
1

skill-builder

Create, evaluate, and improve Claude Skills using templates and scoring rubrics.

Community
Advanced
timurgaleevtimurgaleev
5

canary

Monitor post-deploy web apps for defects and performance regressions.

Community
Advanced
diegocamara89diegocamara89
3

skill-creator

Create and iteratively improve AI skills with evaluation workflows.

Community
Advanced
charly-vibescharly-vibes

extract-prompt

Extract reusable prompt templates from AI conversations by analyzing successful patterns.

Community
Intermediate