skills.rest
kouko
koukoCOMMUNITY
Explore all skills by this author →
GitHub

skill-creator-advance

Create and iteratively improve AI skills with eval-driven development.

6|1|Updated Apr 2, 2026
Search Tags:#automation#benchmarking#packaging#claude-code#skill-authoring#evals#description-optimization
One-click install
npx skills add https://github.com/kouko/monkey-skills --skill skill-creator-advance
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-creator-advance
Source: https://github.com/kouko/monkey-skills/tree/main/dev-workflow/skills/skill-creator-advance
Command: npx skills add https://github.com/kouko/monkey-skills --skill skill-creator-advance

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Help authors reliably create, evaluate, and improve Claude skills by providing a guided, eval-driven development loop that drafts SKILL.md, runs with-skill and baseline evaluations, grades results, detects regressions, and packages distributable .skill artifacts. It reduces manual trial-and-error, surfaces flaky or non-discriminating tests, and improves triggering accuracy via description optimization.

Core Features & Use Cases

  • Iterative Eval Loop: Draft → test (with baseline) → grade → improve → repeat, with support for quick vs full eval paths.
  • Automated Grading & Benchmarking: Spawn grader/comparator/analyzer subagents, aggregate statistics, and generate benchmark reports and HTML visualizations.
  • Description Optimization: Generate trigger evals, run train/test splits, and iterate descriptions using the claude CLI to avoid under-/over-triggering.
  • Packaging & Automation: Validation, packaging to .skill, and helper scripts for report generation and benchmark aggregation.
  • Use Case: A developer converting a multi-step workflow into a robust skill can draft SKILL.md, supply a few realistic test prompts, run parallel evals, inspect grading, fix issues, and produce a packaged skill for distribution.

Quick Start

Ask: "Help me create a new Claude skill — draft SKILL.md, propose test cases, run evals with baselines, grade results, and suggest improvements until tests pass."

Frequently Asked Questions about skill-creator-advance

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

FAQPage Schema
How do I create and iterate AI skills using eval-driven development?▼

Eval-driven AI skill development involves drafting a SKILL.md file, running automated baseline evaluations, grading results, and iterating until tests pass. This loop reduces manual trial-and-error by surfacing flaky tests early.

How does automated benchmarking work for Claude skill authoring?▼

Automated benchmarking for Claude skill authoring works by spawning grader and analyzer subagents to evaluate test prompts, aggregating statistics, and generating benchmark reports with HTML visualizations to detect regressions.

What is the best way to optimize skill descriptions to avoid under-triggering?▼

To optimize skill descriptions and avoid under-triggering, generate trigger evals, run train/test splits, and iterate descriptions using the claude CLI to ensure accurate activation across various prompts.

Can I package distributable .skill artifacts after authoring?▼

Yes, you can package distributable .skill artifacts after authoring by validating the SKILL.md structure and using helper scripts to bundle the final skill for distribution.

Do I need pyyaml installed to run skill evaluations and baselines?▼

Yes, pyyaml is a required dependency to run skill evaluations and baselines, supporting the parsing of configuration files needed to spawn evaluation subagents.

Why does my AI skill fail to trigger correctly on realistic test prompts?▼

An AI skill fails to trigger correctly when descriptions are poorly optimized or test cases are non-discriminating. Running description optimization and iterative grading helps identify and fix these triggering inaccuracies.

Related Skills

Explore More Software Engineering Skills→
Fudan-SMI-labFudan-SMI-lab
39

torch-npu-pyyaml-preinstall

Preinstall PyYAML before torch-npu to prevent transitive dependency failures.

Official
Intermediate
WUIX69WUIX69

eval-harness

Define evaluation frameworks for Claude Code tasks with pass@k metrics.

Community
Advanced
dparedesidparedesi
1

skill-builder

Create, evaluate, and improve Claude Skills using templates and scoring rubrics.

Community
Advanced
tuliosouza99tuliosouza99

skill-creator

Create and iteratively improve SKILL.md-based agent skills with Python eval scripts.

Community
Advanced