skills.rest
AgentTeam-TaichuAI
AgentTeam-TaichuAICOMMUNITY
Explore all skills by this author →
GitHub

skill-creator

Draft, test, and refine reusable agent skills within the ScienceClaw framework.

630|68|Updated Mar 13, 2026
Search Tags:#workflow#documentation#evaluation#benchmark#skill-writing#skill-creator
One-click install
npx skills add https://github.com/AgentTeam-TaichuAI/ScienceClaw --skill skill-creator-agentteam-taichuai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-creator
Source: https://github.com/AgentTeam-TaichuAI/ScienceClaw/tree/main/ScienceClaw/backend/builtin_skills/skill-creator
Command: npx skills add https://github.com/AgentTeam-TaichuAI/ScienceClaw --skill skill-creator-agentteam-taichuai

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill provides a complete, repeatable workflow for creating, testing, and refining reusable agent skills. It helps teams draft SKILL.md, set up eval/benchmark cycles, and engineer improvements in a privacy-safe ScienceClaw environment.

Core Features & Use Cases

  • Draft, validate, and iterate the structure of a Skill unit (SKILL.md, scripts/, references/, assets/) to ensure discoverability and safe execution.
  • Run evaluation and benchmarking cycles to quantify skill performance, iterate on descriptions, and bundle improvements into releases.
  • Use ScienceClaw-specific guidance to maintain security, transparency, and reproducibility across the Skill development lifecycle.

Quick Start

Use this skill to draft a new Skill, run an evaluation loop, and iteratively improve the description and workflow until evaluation results meet your criteria.

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 refine reusable agent skills?▼

To create and refine reusable agent skills, draft a SKILL.md file, run evaluation and benchmarking cycles, and iteratively improve the workflow until performance meets your criteria.

What is the best way to structure a skill for safe execution and discoverability?▼

The best way to structure a skill is to validate and iterate its core components, including SKILL.md, scripts, references, and assets, ensuring it remains discoverable and safe to execute.

How do I set up evaluation and benchmark cycles to quantify skill performance?▼

You set up evaluation and benchmark cycles to quantify skill performance by testing the skill, iterating on descriptions, and bundling improvements into releases within the ScienceClaw framework.

Do I need Python dependencies to draft and test agent skills?▼

Yes, you need Python dependencies like anthropic and pyyaml to support the skill drafting, evaluation, and structured improvement processes within the development workflow.

Can I maintain security and reproducibility across the skill development lifecycle?▼

You can maintain security and reproducibility across the skill development lifecycle by applying ScienceClaw-specific guidance to ensure transparency during skill creation and testing.

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
gelowzzzgelowzzz

skill-creator

Create, validate, and package Claude Skills into .skill files.

Community
Advanced
cslsoftwarescslsoftwares

developer-web-documentation-governance

Govern web kit skill documentation with changelogs, source references, and risk assessments.

Community
Advanced
codeboltaicodeboltai

skill-creator

Create and refine Claude skills with structured SKILL.md templates and packaging workflows.

Official
Intermediate