skill-creator

Create and iterate SKILL.md descriptions with structured eval sets and trigger benchmarking.

37|4|Updated Mar 31, 2026
One-click install
npx skills add https://github.com/MerkyorLynn/Lynn --skill skill-creator-merkyorlynn
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/MerkyorLynn/Lynn/tree/main/skills2set/skill-creator
Command: npx skills add https://github.com/MerkyorLynn/Lynn --skill skill-creator-merkyorlynn

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Skill-creator helps you go from an idea to a high-quality Claude Code skill by guiding the full build-and-validate loop, including writing or editing SKILL.md, creating an eval set, running trigger tests, and iterating on the description until triggering accuracy improves.

Core Features & Use Cases

  • Design skill intent and trigger behavior: clarify what the skill should do, when it should trigger, and what output format to expect.
  • Create/iterate SKILL.md descriptions: write a first draft, then refine the description based on measurable evaluation results.
  • Run trigger evaluations and benchmarking: generate an eval set, execute runs, compare with baselines, and summarize results with variance-style reporting.
  • Use description optimization: automatically search for a better SKILL.md frontmatter description to improve invocation reliability.

Quick Start

Tell the assistant what you want your new skill to do and when it should trigger (e.g., “Create a skill that turns X into Y whenever the user mentions Z”).

Frequently Asked Questions about skill-creator

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

FAQPage Schema
How do I improve Claude Code skill triggering accuracy for my custom SKILL.md descriptions?

Improve Claude Code skill triggering accuracy by creating structured eval sets, running automated trigger benchmarking, grading pass rates, and iteratively refining SKILL.md descriptions using train/test holdout and benchmark aggregation.

What's the best way to benchmark and evaluate AI skill invocation reliability?

Benchmark AI skill invocation reliability by generating an eval set, executing trigger runs, comparing results with baselines, and summarizing findings with variance-style reporting to measure and optimize description tuning.

Can I optimize SKILL.md frontmatter descriptions automatically to fix incorrect skill invocation?

You can automatically search for a better SKILL.md frontmatter description to fix incorrect skill invocation using the description optimization feature to improve trigger reliability.

Do I need Anthropic and pyyaml dependencies to run automated trigger evals for Claude Code?

Yes, you need Anthropic and pyyaml dependencies installed in your environment to run automated trigger evals, execute benchmarking runs, and process SKILL.md frontmatter formatting.

How does description optimization handle overfitting when refining SKILL.md trigger text?

Description optimization handles overfitting when refining SKILL.md trigger text by applying train/test holdout splits during eval generation, ensuring trigger improvements generalize across unseen benchmark prompts.