skill-creator

Create, evaluate, and refine Claude skills with SKILL.md and eval prompts.

1|Updated Nov 20, 2025
One-click install
npx skills add https://github.com/tony-zhelonkin/SciAgent-toolkit --skill skill-creator-tony-zhelonkin
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/tony-zhelonkin/SciAgent-toolkit/tree/main/skills/skill-creator
Command: npx skills add https://github.com/tony-zhelonkin/SciAgent-toolkit --skill skill-creator-tony-zhelonkin

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates the end-to-end process of creating, evaluating, and refining Claude-based skills, turning experimentation into repeatable, measurable workflow.

Core Features & Use Cases

  • Guided process to draft SKILL.md, design eval prompts, and run iterative improvements
  • Built-in support for train/test evaluation, benchmarking, and result sharing
  • Tools for packaging, validation, and description optimization to scale skill deployment
  • Clear guidance on writing accessible, robust skills that generalize across domains

Quick Start

Use the skill-creator workflow to draft a new skill, run initial evals on a sample prompt, and iterate until you achieve stable performance.

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 iterate Claude skills with a repeatable workflow?

Creating and iterating Claude skills involves a structured workflow to draft SKILL.md files, design evaluation prompts, run train/test benchmarks, and refine outputs until performance stabilizes.

What is the process for evaluating prompt engineering skills?

Evaluating prompt engineering skills uses built-in train/test prompts and benchmarking across sample inputs to measure performance and guide iterative refinement.

How do I package and validate skills for scalable deployment?

Packaging and validating skills for scalable deployment involves specifying required frontmatter, defining error-handling patterns, and applying validation tools to ensure robust generalization across domains.

Do I need to define error-handling patterns to build robust Claude skills?

Defining error-handling patterns is required to build robust Claude skills, ensuring stable performance and reliable generalization across different contexts and edge cases.

When should I use a structured skill creation workflow instead of manual experimentation?

Use a structured skill creation workflow instead of manual experimentation when you need to convert iterative prompt testing into measurable, repeatable processes for scalable skill deployment.