skills.rest
eddie-292
eddie-292COMMUNITY
Explore all skills by this author →
GitHub

skill-creator

Create, test, and iterate on Claude skills with evaluation workflows.

52|8|Updated Mar 12, 2026
Search Tags:#prompt engineering#benchmarking#claude#skill authoring#evals#description optimization
One-click install
npx skills add https://github.com/eddie-292/grapwork --skill skill-creator-eddie-292
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/eddie-292/grapwork/tree/main/frontend/skills/examples/skill-creator
Command: npx skills add https://github.com/eddie-292/grapwork --skill skill-creator-eddie-292

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill helps you turn an idea into a reliable skill, then improve it through testing and feedback instead of guessing. It is designed for drafting new skills, editing existing ones, measuring trigger quality, and tightening descriptions so the right tasks activate the skill more often.

Core Features & Use Cases

  • Skill Drafting: Shape a clear SKILL.md with the right frontmatter, instructions, and supporting resources.
  • Evaluation Workflow: Create realistic test prompts, run comparisons, and review pass/fail results to see where the skill succeeds or misses.
  • Iteration and Optimization: Refine the skill body, improve trigger wording, and use benchmarks to spot weak spots or overfitting.
  • Practical Use Cases: Building a new workflow skill from scratch, repairing an underperforming skill, or tuning a skill description so it triggers more consistently on real user requests.

Quick Start

Tell the skill creator what you want the skill to do, and it will help you draft the skill, prepare test prompts, run evaluations, and revise the description based on the results.

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 test Claude skills from scratch?▼

To create Claude skills, you draft a SKILL.md with frontmatter and instructions, then run evaluation prompts to compare pass/fail results and iteratively refine the skill body. This workflow ensures your new skill is tested and validated before release.

What is skill trigger optimization and how does benchmarking improve it?▼

Skill trigger optimization uses benchmarking to measure how consistently a skill activates on real user requests. By running evaluation workflows and analyzing pass/fail results, you can tighten descriptions and refine trigger wording to improve accuracy.

Can I use evaluation prompts to fix an underperforming skill?▼

Yes, you can repair an underperforming skill by designing realistic test prompts and running comparative evaluations. Benchmark-driven feedback spots weak spots or overfitting, allowing you to revise instructions and improve trigger accuracy.

Do I need pyyaml and Claude to run skill evaluation workflows?▼

Yes, skill authoring and evaluation workflows require Claude and pyyaml dependencies to validate skill metadata, run iterative testing scripts, and keep instructions, resources, and descriptions aligned throughout the process.

What's the best way to draft a SKILL.md with the right frontmatter?▼

The best way to draft a SKILL.md is to define the core task, then generate the appropriate frontmatter, instructions, and supporting resources. Iterative evaluation and validation scripts ensure the metadata stays aligned with the skill's actual behavior.

Related Skills

Explore More Software Engineering Skills→
ibnuhalimmibnuhalimm

Architect

Gather project context and produce a detailed architecture plan.

Community
Advanced
pipixia-labspipixia-labs
38

skill-creator

Guide creation of SKILL.md files and coordinate evals, benchmarks, and iteration.

Official
Advanced
anymouschinaanymouschina
523

skill-creator

Draft, refine, and evaluate Claude Skills with SKILL.md and evals.

Community
Advanced
beyonaibeyonai
45

skill-optimizer

Optimize agent skill instructions through a validation-gated feedback loop.

Community
Advanced