skills.rest
thevibeworks
thevibeworksOFFICIAL
Explore all skills by this author →
WebsiteGitHub

skill-creator

Create and iteratively improve Claude skills through evaluation loops.

28|5|Updated Jun 24, 2025
Search Tags:#prompt engineering#skill creation#benchmarking#evaluation#iteration loop#skill triggering
One-click install
npx skills add https://github.com/thevibeworks/claude-code-docs --skill skill-creator-thevibeworks
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/thevibeworks/claude-code-docs/tree/main/content/github/claude-plugins-official/plugins/skill-creator/skills/skill-creator
Command: npx skills add https://github.com/thevibeworks/claude-code-docs --skill skill-creator-thevibeworks

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The skill-creator helps you turn an unclear idea for a Claude/Agent Skill into a working, well-tested skill that triggers reliably and performs well on real prompts.

Core Features & Use Cases

  • Skill draft to iteration loop: Captures intent, drafts a skill, then iterates based on qualitative and quantitative evaluation results.
  • Evaluation harness setup: Creates realistic test prompts, runs them with and without the skill, and uses a review viewer to compare behavior and outcomes.
  • Description optimization: Improves triggering accuracy by running a trigger-eval optimization loop against your skill description.
  • Use cases: Create a skill from scratch, edit an existing skill, optimize description triggering, and benchmark skill performance with variance analysis.

Quick Start

Ask: "Turn my workflow for [X] into a Claude skill, and help me test and improve it until it reliably triggers and performs well."

Frequently Asked Questions about skill-creator

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

FAQPage Schema
How do I create a Claude skill from scratch and make sure it triggers reliably?▼

To create a Claude skill, you capture user intent and draft SKILL.md instructions, then run an evaluation loop with test prompts to iteratively improve performance and optimize the description for accurate triggering.

What is the best way to benchmark an existing AI skill's performance?▼

The best way to benchmark skill performance is by running a set of realistic test prompts with and without the skill, then using a review viewer to compare outcomes and conduct variance analysis on the results.

How do I improve my skill description so it triggers accurately across diverse prompts?▼

You improve skill description triggering by running a trigger-eval optimization loop that tests the description against diverse prompts and iteratively refines it based on quantitative evaluation evidence.

Can I edit an existing skill without rebuilding the entire SKILL.md from scratch?▼

Yes, you can edit an existing skill by capturing your revised intent, updating the SKILL.md instructions, and running evaluation loops to test the updated behavior against baseline performance.

Does skill creation require setting up an evaluation harness before testing?▼

Yes, skill creation requires an evaluation harness setup to generate realistic test prompts, execute them with and without the skill, and compare behavioral outcomes to establish a reliable baseline.

Related Skills

Explore More Software Engineering Skills→
Crispy13Crispy13

skill-creator-copilot

Draft SKILL.md files with frontmatter, instructions, and test workflows.

Community
Advanced
steadyfallsteadyfall

canary

Monitor live applications for console errors and performance regressions after deployment.

Community
Advanced
selcukyucelselcukyucel
1

eval-suite

Generate structured evaluation datasets with binary rubric scoring for AI outputs.

Community
Advanced
yasuwrldhyperyasuwrldhyper

skill-creator

Create and refine AI skills with SKILL.md metadata and benchmark testing.

Community
Advanced