skill-creator

Automate the creation, evaluation, and iteration of Claude skills with SKILL.md drafting and eval loops.

Updated Oct 24, 2024
One-click install
npx skills add https://github.com/tuannguyen151/nestjs-clean-architecture --skill skill-creator-tuannguyen151
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/tuannguyen151/nestjs-clean-architecture/tree/main/.claude/skills/skill-creator
Command: npx skills add https://github.com/tuannguyen151/nestjs-clean-architecture --skill skill-creator-tuannguyen151

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Automates the end-to-end process of creating, evaluating, and refining Claude skills from scratch, helping teams capture intent, author SKILL.md, run iterative eval loops, and optimize triggering descriptions without manual drift.

Core Features & Use Cases

  • Draft and update SKILL.md with guided prompts and sections for problem, features, and quick-start usage.
  • Run trigger evaluations across diverse prompts, collect results, and generate evaluative benchmarks.
  • Iterate on skill descriptions using automated improvement loops to increase triggering accuracy and reliability.
  • Compare iterations using blinded or post-hoc analysis to identify effective changes.
  • Package skills into distributable formats and manage references/scripts/assets for on-demand execution.

Quick Start

Run the loop once by pointing this skill at an eval-set and your skill folder; it will automatically evaluate, improve, and re-evaluate until optimization converges.

Frequently Asked Questions about skill-creator

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

FAQPage Schema
How do I automate Claude skill creation and evaluation?

Automate Claude skill creation by drafting SKILL.md files, running iterative evaluation loops across diverse prompts, and optimizing trigger descriptions for accuracy without manual drift. It packages references, scripts, and assets for on-demand execution and testing.

How do I optimize trigger descriptions for Claude skills?

Optimize trigger descriptions by running automated improvement loops that evaluate skills against an eval-set, collect results, and generate benchmarks to increase triggering accuracy and reliability until convergence is reached.

What is the best way to iterate on SKILL.md files?

Iterate on SKILL.md files by comparing iterations using blinded or post-hoc analysis to identify effective changes, capturing intent, and updating sections for problems, features, and quick-start usage automatically.

Do I need pyyaml and anthropic dependencies to run skill automation?

Yes, you need the anthropic and pyyaml dependencies to run skill automation. They support the underlying evaluation loops and configuration parsing required to draft, evaluate, and refine Claude skills.

Can I package Claude skills into distributable formats with references and scripts?

Package Claude skills into distributable formats by coordinating optional resources including scripts, references, and assets. This manages on-demand execution and testing alongside the core SKILL.md file.