skill-creator

Creates, evaluates, and iterates Claude skills with benchmark-driven improvements.

Updated Mar 19, 2026

Search Tags:#automation #evaluation #iteration #benchmark #skill #prompt

Ready to run

npx skills add https://github.com/stabilefrisur/panmetis --skill skill-creator-stabilefrisur

💬 Prompt for ChatGPT / Claude Chat▼

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

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires 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 iterating Claude skills, reducing guesswork and accelerating improvement cycles.

Core Features & Use Cases

Draft new skills from user intent and structure their frontmatter and body.
Run trigger evaluations and benchmarks to quantify triggering accuracy, latency, and coverage.
Iterate skill descriptions, scripts, and references based on evaluation results and user feedback.
Bundle optional resources (scripts, references, assets) for on-demand use.

Quick Start

Start by creating a skill folder with SKILL.md and run the evaluation loop to begin refining its description and behavior.

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 frontmatter and body from user intent, then run trigger evaluations and benchmarks to quantify triggering accuracy, latency, and coverage for iterative improvements.

What's the best way to iterate skill descriptions to improve trigger accuracy?▼

Iterate skill descriptions by running trigger evaluations and benchmarks, then refining descriptions, scripts, and references based on evaluation results and user feedback to achieve history-aware improvements.

How does benchmarking Claude skills work for reproducibility?▼

Benchmarking Claude skills works by running trigger evaluations that quantify accuracy, latency, and coverage, ensuring reproducibility and safe, auditable outputs across history-aware improvement cycles.

Do I need PyYAML to create and test Claude skills?▼

Yes, you need PyYAML installed to create and test Claude skills, as it is the required dependency for parsing and structuring skill frontmatter during the drafting and iteration process.

Can I bundle scripts and references for on-demand skill use?▼

Yes, you can bundle optional resources including scripts, references, and assets for on-demand use, ensuring safe and auditable outputs while maintaining reproducibility across skill iterations.

Why does my Claude skill trigger evaluation not produce reproducible results?▼

Trigger evaluation results may lack reproducibility if skill descriptions and references are not iterated using history-aware improvements and benchmarking to ensure safe, auditable outputs across evaluation cycles.