skill-creator

Draft Claude skill instructions, generate evals, and run comparisons.

1|Updated Apr 16, 2026
One-click install
npx skills add https://github.com/victorjoshi9/content--1- --skill skill-creator-victorjoshi9
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/victorjoshi9/content--1-/tree/main/.claude/plugins/marketplaces/claude-plugins-official/plugins/skill-creator/skills/skill-creator
Command: npx skills add https://github.com/victorjoshi9/content--1- --skill skill-creator-victorjoshi9

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Skill Creator helps you turn an idea into a working Claude skill, then improve it with real tests instead of guesswork. It is designed for the full lifecycle: drafting the skill, checking whether it triggers for the right requests, and tightening the instructions until performance is reliable.

Core Features & Use Cases

  • Draft new skills: Start from a user goal and shape it into a clear SKILL.md with strong triggering language.
  • Iterate with evaluations: Create test prompts, run comparisons, review outputs, and revise the skill based on what actually happened.
  • Measure performance: Benchmark different versions, compare variance, and spot weak assertions or flaky behavior.
  • Improve triggering accuracy: Refine the description so the skill activates for the right kinds of requests and stays quiet for near misses.
  • Example use case: You want a skill that helps teammates summarize meeting notes from docs and spreadsheets; use this skill to write the first draft, test it, and sharpen the description until it triggers consistently.

Quick Start

Use the skill creator to draft a new skill for your workflow, test it with realistic prompts, and keep refining the description until it behaves the way you want.

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 Claude skills with prompt evaluation?

You can create and test Claude skills by drafting a SKILL.md, generating test prompts, running benchmarking comparisons, and iterating on instructions based on real output evaluations. This end-to-end workflow replaces guesswork with measured performance validation.

What is the best way to improve skill triggering accuracy for prompt requests?

To improve skill triggering accuracy, refine the skill description so it activates for the right requests and stays quiet for near misses. Iterate on the description language and run trigger tuning evaluations to measure activation behavior.

How do I benchmark different versions of a skill to compare variance?

Benchmark different skill versions by running comparison evaluations across test prompts, reviewing outputs, and measuring variance to spot weak assertions or flaky behavior. This performance measurement workflow validates which version behaves reliably.

Can I use this skill creator for document and content task workflows?

Yes, the skill creator supports end-to-end workflows for new skill creation, refinement, benchmarking, and packaging across code, document, and content tasks. You can draft a skill for summarizing meeting notes or similar document workflows.

Do I need Python and PyYAML installed to build and refine skills?

Yes, you need the Anthropic and PyYAML dependencies installed in your environment to run the scripts, references, and review artifacts required for skill creation, evaluation, and benchmarking workflows.