skill-creator

Create workflows for drafting, testing, and evaluating OpenCode skills.

2|Updated Oct 12, 2011

Search Tags:#evaluation #iteration #benchmark #opencode #skill-creation

Ready to run

npx skills add https://github.com/aming/dotfiles --skill skill-creator-aming

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/aming/dotfiles/tree/main/agents/dot-agents/skills/skill-creator
Command: npx skills add https://github.com/aming/dotfiles --skill skill-creator-aming

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

The Skill Creator provides a structured workflow to design, test, and iteratively improve OpenCode skills, replacing ad-hoc experimentation with measurable progress.

Core Features & Use Cases

Draft new skills from user ideas and map them to evaluation plans.
Run end-to-end eval loops that include training/test splits, prompts, and benchmarking data.
Automatically propose improvements to skill descriptions based on eval results, history, and benchmarks.
Package polished skills for distribution and reuse.

Quick Start

Draft a new skill, run an initial eval loop with sample prompts, and iteratively refine the description until results stabilize.

Frequently Asked Questions about skill-creator

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

FAQPage Schema

How do I benchmark an OpenCode skill during development?▼

To benchmark an OpenCode skill, you run deterministic evaluation loops using test prompts, train/test splits, and benchmark reporting to measure iterative progress and stabilize results.

What is the best way to structure a workflow for skill creation?▼

The best way to structure skill creation is to map user ideas to evaluation plans, run end-to-end eval loops, propose description improvements based on results, and package the polished skill.

Do I need YAML and Anthropic dependencies to build skills?▼

Yes, you need YAML and Anthropic dependencies to build and refine OpenCode skills, as they support the underlying workflow scripts, evaluation criteria, and asset components required for skill processing.

Can I automatically refine a skill description based on evaluation history?▼

Yes, you can automatically refine a skill description by analyzing evaluation results, history, and benchmark data to propose actionable improvements until the skill's performance stabilizes.

How does iterative testing improve skill packaging for distribution?▼

Iterative testing improves packaging by applying deterministic evaluation through train/test splits and benchmark reporting, ensuring the skill description is polished and ready for distribution.