skill-creator

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

2|Updated Mar 21, 2026
One-click install
npx skills add https://github.com/ghoshsam/claude-code-skills --skill skill-creator-ghoshsam
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/ghoshsam/claude-code-skills/tree/main/skills/skill-creator
Command: npx skills add https://github.com/ghoshsam/claude-code-skills --skill skill-creator-ghoshsam

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve? Writing a skill that triggers reliably and produces good outputs is hard to get right on the first try. This Skill guides the full lifecycle of skill development — drafting, testing against real prompts, benchmarking with quantitative assertions, collecting human feedback, and optimizing the description for accurate triggering. ## Core Features & Use Cases - Guided Skill Authoring: Interviews the user about intent, edge cases, and output formats, then writes a well-structured SKILL.md following progressive disclosure and writing-style best practices. - Evaluation & Benchmarking: Runs test prompts with and without the skill in parallel, grades outputs against assertions, aggregates pass rates, timing, and token usage, and renders results in an interactive HTML review viewer. - Description Optimization: Generates realistic trigger/no-trigger eval queries and runs an automated optimization loop to improve how reliably the skill gets invoked. - Use Case: You want to build a skill that converts meeting notes into structured action items. This Skill helps you draft it, creates test cases, runs them, shows you the outputs side-by-side with baselines, and iterates until the results satisfy you. ## Quick Start Help me create a new skill that turns my weekly meeting notes into a structured list of action items with owners and deadlines.

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 edge cases and output formats, drafts a SKILL.md with proper frontmatter, then helps you test and refine it through evaluation loops.

How do I test whether my skill actually works?

Define 2-3 realistic test prompts, then run them with and without the skill in parallel. Outputs are graded against assertions, aggregated into a benchmark with pass rates and timing, and displayed in an HTML review viewer for qualitative feedback.

How can I improve my skill description so it triggers more reliably?

Generate 20 realistic trigger and no-trigger eval queries, then run the automated optimization loop. It evaluates description variants on train and held-out test splits across up to 5 iterations and returns the best-scoring description.

What makes a good SKILL.md file?

A good SKILL.md has a name and a pushy, context-rich description in the frontmatter, keeps the body under 500 lines, uses imperative instructions that explain why, and offloads detail to references, scripts, and assets directories.

Does skill evaluation work on Claude.ai without subagents?

Yes, but with limitations. Without subagents you run test prompts sequentially yourself, skip baseline comparisons and quantitative benchmarking, and collect feedback inline in the conversation instead of through the browser viewer.