skills.rest
NGTPTE
NGTPTECOMMUNITY
Explore all skills by this author →
GitHub

skill-creator

Create, test, and improve reusable Claude Code skills from a draft SKILL.md.

Updated Apr 8, 2026
Search Tags:#claude code#benchmarking#skills#packaging#evals#description tuning
One-click install
npx skills add https://github.com/NGTPTE/TuriX-MarketPlace --skill skill-creator-ngtpte
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/NGTPTE/TuriX-MarketPlace/tree/main/plugins/skill-creator/skills/skill-creator
Command: npx skills add https://github.com/NGTPTE/TuriX-MarketPlace --skill skill-creator-ngtpte

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires anthropic, pyyaml, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill helps you turn an idea for a Claude Code skill into a validated, well-triggering package, and then improve an existing skill with objective evals instead of guesswork.

Core Features & Use Cases

  • Draft and refine SKILL.md content for new or existing skills.
  • Generate realistic trigger evals, run baseline comparisons, and review pass rates.
  • Iterate on descriptions when a skill under-triggers or over-triggers, then package the final folder for sharing.

Quick Start

Tell the skill creator what behavior you want the skill to support, and it will help you draft the skill, build evals, run tests, and tighten the description until it triggers reliably.

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 test reusable Claude Code skills?▼

To create Claude Code skills, draft a SKILL.md file defining the desired behavior, then generate trigger evals, run baseline comparisons, and iterate on the description until the skill reliably triggers.

What is the best way to fix a Claude Code skill that over-triggers or under-triggers?▼

To fix skill triggering issues, run trigger evals to benchmark pass rates, then apply description tuning to refine the SKILL.md content until the skill activates correctly for the intended inputs.

Can I benchmark and grade evals for an existing Claude Code skill?▼

Yes, you can benchmark and grade evals for an existing skill by running baseline comparisons and reviewing pass rates to objectively measure performance and identify areas for iterative optimization.

How do I package a Claude Code skill for sharing after refining it?▼

To package a Claude Code skill for sharing, finalize the SKILL.md content and bundled scripts, then use the packaging workflow to assemble the finished folder structure with its references and review assets.

Do I need Anthropic and pyyaml dependencies to build Claude Code skills?▼

Yes, the skill creation, eval generation, and benchmarking workflows require the Anthropic and pyyaml dependencies to function properly alongside the bundled scripts and review assets.

Related Skills

Explore More Software Engineering Skills→
Shockwave2kShockwave2k

skill-creator

Draft SKILL.md files and run evaluation loops to measure trigger accuracy.

Community
Advanced
epologeeepologee
3

export-skill

Sanitize, translate, package, and share Claude skills sequentially.

Community
Intermediate
gesmith0606gesmith0606

skill-creator

Create, evaluate, and iteratively improve SKILL.md descriptions with Python scripts.

Community
Advanced
dsteven12dsteven12
1

skill-creator

Create, test, and improve Cowork skills through automated evals and iteration.

Community
Advanced