一键导入
create-skill
Help create a new AI agent skill through collaborative Q&A
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Help create a new AI agent skill through collaborative Q&A
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Analyze R/Bioconductor package structure to extract key information about its purpose, exports, and characteristics
Verify that waldronlab/ai-agent-skills are available and help users find the right skill
Update existing AI instruction documentation when an R/Bioconductor package changes
Validate that a skill conforms to the ai-agent-skills repository standards
Create complete AI instruction documentation for an R/Bioconductor package in .github/instructions/
Automate documentation updates to SKILLS.md after creating or modifying a skill
| name | create-skill |
| description | Help create a new AI agent skill through collaborative Q&A |
| version | 1.3.0 |
| category | meta |
| tags | ["meta","infrastructure","skill-creation"] |
| author | waldronlab |
Help create new AI agent skills through collaborative Q&A. Guides from rough ideas to structured skill files, focusing on intent over perfect formatting.
Invoke this skill when you want to create a new skill:
Start with open-ended questions:
Existing domains: meta, r-packages, metagenomics, statistical-methods
If uncertain, suggest the closest match. Convert purpose to kebab-case for the filename: "analyze R package" → analyze-r-package.
Ask the user to walk through the steps. Capture major steps, key decisions, required inputs, expected outputs, and strategic code snippets (as guardrails). Help clarify and break down complex steps.
Create a well-structured skill file following AGENTS.md and SKILL_STANDARD.md.
Do not duplicate SSOT guidelines; use references instead.
Fill in what's clear, use TODO markers for uncertainties, and prioritize clarity.
Present a summary to the user, invite feedback ("Did I capture your intent correctly?"), and offer to iterate.
Before updating repository documentation, invoke validate-skill on the new file. Fix any CRITICAL issues. Only proceed to step 7 if validation passes.
Invoke document-skill by saying "Document the [skill-name] skill". It will automatically update SKILLS.md. Confirm completion with the user and suggest they test the skill.
Generate a markdown file saved to skills/[skill-name]/SKILL.md. See SKILL_STANDARD.md for the required template and format.
User: "I want to make a skill that helps validate R package documentation"
Agent: Engages in Q&A to understand documentation aspects, timing, and actions.
Result: skills/validate-r-documentation/SKILL.md
validate-skill before document-skillRelated: See SKILL_STANDARD.md for technical format details. See AGENTS.md for how agent discovery works.