skill-creator

Create and iteratively improve Claude Code skills with trigger evals.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/cwinux/skillflaw --skill skill-creator-cwinux
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/cwinux/skillflaw/tree/main/skill-creator
Command: npx skills add https://github.com/cwinux/skillflaw --skill skill-creator-cwinux

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill helps you rapidly create new Claude Code skills or improve existing ones by turning vague goals into a working skill loop that includes drafting, test creation, evaluation runs, quantitative review, and iterative refinement.

Core Features & Use Cases

  • Skill design from scratch: Convert an intent into a skill draft by eliciting trigger conditions and expected outputs.
  • Evaluation-driven iteration: Generate and run trigger-evaluation test prompts, then refine based on measured failure modes.
  • Performance measurement and optimization: Benchmark triggering accuracy (and associated costs) and optionally optimize the skill description to improve how reliably Claude selects it.
  • Feedback-centered workflow: Guide you through reviewing outputs and incorporating structured feedback until the skill performs as desired.

Quick Start

Tell the skill you want to build a new skill for doing X, and include when it should trigger and what the output should look like.

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 from scratch?

Creating Claude Code skills involves drafting deterministic skill descriptions, generating structured trigger evaluation tests, and running an iteration loop to measure and refine triggering accuracy until performance goals are met.

What is trigger optimization for agent workflows?

Trigger optimization for agent workflows refines skill descriptions through evaluation loops to measure failure modes, ensuring Claude accurately and reliably selects the intended skill during task execution.

How do I improve an existing Claude Code skill's triggering accuracy?

Improving an existing skill requires running trigger-evaluation test prompts to identify failure modes, then applying optional description optimization based on quantitative review to enhance invocation reliability.

Can I run evals to measure skill quality and performance variance?

Yes, you can run evals to measure skill quality by generating structured test prompts that benchmark triggering accuracy and calculate associated costs, providing variance-aware evaluation of skill performance.

What do I need to start authoring a new skill for Claude Code?

To start authoring a skill, define the intent, specify when the skill should trigger, and describe the expected output to initiate the drafting, test generation, and evaluation workflow.