skill-creator

Create and iteratively improve AI skills with SKILL.md and evaluation workflows.

Updated Apr 12, 2026

Search Tags:#benchmarking #packaging #scripts #claude-cli #skill-creator #evals #description-optimization

Ready to run

npx skills add https://github.com/MaxServDev/DriClaw-Evolution-local --skill skill-creator-maxservdev

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/MaxServDev/DriClaw-Evolution-local/tree/main/.agents/skills/skill-creator
Command: npx skills add https://github.com/MaxServDev/DriClaw-Evolution-local --skill skill-creator-maxservdev

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Helps developers and product teams turn conversational workflows and repeatable procedures into robust, testable Claude Code skills, reducing manual drafting, ad-hoc testing, and fragile deployments.

Core Features & Use Cases

Draft & Authoring: Guide authors to capture intent, trigger conditions, and structured SKILL.md content for reliable invocation.
Eval & Benchmarking: Produce eval sets, run with-skill and baseline comparisons, grade outputs, aggregate benchmark statistics, and surface high-variance or nondiscriminating assertions.
Iteration & Packaging: Automate improvement loops that refine descriptions, rerun tests, and package validated skills for distribution.
Use Case: Turn a user's multi-step data-extraction workflow into a repeatable skill, run 3+ eval iterations, analyze pass rates, and produce a packaged .skill file.

Quick Start

Run the skill-creator to draft or edit a SKILL.md, create test prompts, run the eval loop, and package the skill when you're ready.

Frequently Asked Questions about skill-creator

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

FAQPage Schema

How do I create and benchmark AI skills for Claude?▼

To create and benchmark AI skills, you draft a SKILL.md, generate test prompts, run baseline and with-skill comparisons, grade outputs, and aggregate benchmark statistics to evaluate performance. This skill automates that entire evaluation workflow.

Do I need the claude CLI and Python to run skill evaluation scripts?▼

Yes, you need the claude CLI and Python environment with pyyaml installed to run the evaluation scripts and description optimization loops. These dependencies support local benchmarking utilities and packaging tools.

What's the best way to automate AI skill description optimization?▼

The best way to automate description optimization is running iterative improvement loops that refine SKILL.md descriptions, rerun test prompts, and analyze pass rates. This skill uses the claude CLI to optimize descriptions based on benchmark results.

How does grading and benchmarking work for conversational AI workflows?▼

Grading and benchmarking works by producing eval sets, running with-skill and baseline comparisons, grading outputs against assertions, and aggregating statistics to surface high-variance or nondiscriminating test cases for iterative skill improvement.

Can I package validated Claude skills for distribution after benchmarking?▼

Yes, you can package validated skills for distribution after benchmarking by running the packaging tools that compile the tested SKILL.md, scripts, references, and assets into a distributable .skill file.

Why are my AI skill eval assertions returning high variance across test runs?▼

High variance in eval assertions occurs when test prompts lack specificity or the SKILL.md trigger conditions are ambiguous. This skill surfaces nondiscriminating assertions during benchmark aggregation to help refine and stabilize outputs.