skill-creator

Create, evaluate, and iteratively improve Claude skills with benchmarked test runs.

1|Updated May 10, 2026
One-click install
npx skills add https://github.com/Tgoldi/claude-skills --skill skill-creator-tgoldi
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/Tgoldi/claude-skills/tree/main/skill-creator
Command: npx skills add https://github.com/Tgoldi/claude-skills --skill skill-creator-tgoldi

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve? Writing a skill that reliably triggers and performs well is hard to do by intuition alone. This Skill guides you through drafting a SKILL.md, building test cases, running with-skill versus baseline comparisons, and optimizing the description so the skill triggers when it should. ## Core Features & Use Cases - Guided Skill Authoring: Interviews you about intent, trigger contexts, and output formats, then writes a well-structured SKILL.md following progressive disclosure and writing-style best practices. - Evaluation & Benchmarking: Spawns parallel with-skill and baseline runs, grades outputs against assertions, and aggregates pass rates, timing, and token usage into benchmark reports with an interactive HTML review viewer. - Description Optimization: Generates realistic trigger/no-trigger eval queries, reviews them with you in an editable HTML page, and runs an automated train/test optimization loop to improve triggering accuracy. - Use Case: You want a skill that formats weekly sales spreadsheets. This Skill helps you draft it, creates three realistic test prompts, runs them with and without the skill, shows you side-by-side results in a browser viewer, and iterates until the outputs pass your review. ## Quick Start Help me create a new skill that converts meeting notes into structured action items, then test and refine it with me.

Frequently Asked Questions about skill-creator

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

FAQPage Schema
How do I create a new Claude skill from scratch?

Start by describing what the skill should do and when it should trigger. The skill-creator interviews you about intent and output formats, drafts a SKILL.md with proper frontmatter, then helps you build test cases and iterate based on evaluation results.

How do I test whether my skill actually improves Claude's output?

Run each test prompt twice in parallel: once with the skill and once without as a baseline. The skill-creator grades outputs against assertions, aggregates pass rates and timing into a benchmark.json, and opens an HTML viewer for side-by-side review.

How do I optimize a skill description for better triggering?

Generate about 20 realistic should-trigger and should-not-trigger queries, review them in the provided HTML editor, then run the automated optimization loop. It evaluates description variants on train and held-out test splits and returns the best-scoring description.

Does skill-creator work on Claude.ai without subagents?

Yes, but with adaptations. Without subagents you run test cases sequentially yourself, skip baseline comparisons and quantitative benchmarking, and present results inline instead of using the browser viewer. Description optimization requires the claude CLI and is skipped.

What makes a good skill description for triggering?

Include both what the skill does and specific contexts for when to use it, since all triggering decisions come from the description. Slightly assertive phrasing helps counteract undertriggering, and eval queries should be substantive multi-step tasks rather than trivial one-liners.