skill-creator

Coordinate end-to-end creation, evaluation, and packaging of Claude Skills.

Updated Mar 28, 2026

Search Tags:#automation #claude #evaluation #iteration #skill

Ready to run

npx skills add https://github.com/sandalsoft/agent-skills --skill skill-creator-sandalsoft

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/sandalsoft/agent-skills/tree/main/skill-creator
Command: npx skills add https://github.com/sandalsoft/agent-skills --skill skill-creator-sandalsoft

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill provides a complete, repeatable workflow for creating new Claude Skills, evaluating their triggering, iterating on feedback, and packaging for distribution.

Core Features & Use Cases

End-to-end skill creation workflow: draft a skill, design evaluation prompts, run trigger evaluations, and iterate based on results.
Integrated evaluation framework: trigger evaluations, timing, and benchmarking scaffolds to gauge robustness and discovery.
Self-improvement and packaging support: versioning, changelogs, and a packaging workflow to share skills.

Quick Start

Use this skill to scaffold a new skill from a conversation, run evals, and iteratively refine triggering and description until you achieve solid 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 evaluate Claude Skills to improve their triggering reliability?▼

You can create and evaluate Claude Skills by drafting a SKILL.md file, running trigger evaluations, collecting benchmarking feedback, and iterating on the description to achieve higher reliability and discoverability in Claude's triggering system.

What is the best way to build an end-to-end workflow for Claude Skill automation?▼

The best way to build an end-to-end workflow for Claude Skill automation is to use a coordinated framework that scaffolds the skill from a conversation, designs evaluation prompts, and iteratively refines triggering based on performance results.

Does the skill creation workflow support versioning and packaging for distribution?▼

Yes, the skill creation workflow supports versioning and packaging for distribution by generating changelogs and utilizing a packaging workflow to ensure skills can be easily shared and updated over time.

Can I run trigger evaluations and timing benchmarks for drafted Claude Skills?▼

Yes, you can run trigger evaluations and timing benchmarks for drafted Claude Skills using the integrated evaluation framework, which provides scaffolds to gauge robustness and discovery before finalizing the skill.

Do I need pyyaml installed to scaffold and iterate on Claude Skills?▼

Yes, you need pyyaml installed, as it is a required dependency for running the scripts that coordinate skill creation, trigger evaluation, and the iterative improvement workflow.