Skip to content
Back to search

claude-meta-skill

SkillWorks with Claude CodeGitHub

A curated collection of reusable skills for Claude Code. Enhance Claude's capabilities with ready-to-use skill modules including comprehensive guides, templates, and best practices for creating your own skills.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/yyh211--claude-meta-skill/badge.svg)](/listings/yyh211--claude-meta-skill)

HTML

<a href="/listings/yyh211--claude-meta-skill"><img src="/api/listings/yyh211--claude-meta-skill/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

1. **Navigate to your project**

   cd /path/to/your/project

2. **Create the skills directory** (if not exists)

   mkdir -p .claude/skills

3. **Copy the skill you want**

Example - Adding the skill creation guide:

   # clone this repository if you haven't already
   git clone https://github.com/YYH211/Claude-meta-skill.git

   # For Chinese version
   cp -r /Claude-meta-skill/create-skill-file .claude/skills/

   # OR for English version
   cp -r /Claude-meta-skill/create-skill-file-EN .claude/skills/

4. **Verify installation**

   ls .claude/skills/

You should see your copied skill directory:

   .claude/skills/
   ├── create-skill-file/    (or create-skill-file-EN/)
   │   ├── SKILL.md
   │   ├── templates/
   │   └── examples/

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 621 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # Claude Skills Repository
  2. A curated collection of reusable skills for Claude Code. Pick the skills you need and add them to your project to enhance Claude's capabilities.
  3. [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
  4. [![Claude Code Skill](https://img.shields.io/badge/Claude%20Code-Skill-blue)](https://docs.anthropic.com/en/docs/agents-and-tools/claude-code/skills)
  5. ## 🎯 What is This?
  6. This repository provides ready-to-use skills that extend Claude Code's functionality. Each skill is a self-contained module that teaches Claude how to perform specific tasks or follow particular workflows in your projects.
  7. ## 📦 Available Skills
  8. | Skill Name | Description | Source | Installation |
  9. |----------------------------------------------------|-------------------------------------------------------------------------------------------------------|--------|--------------|
  10. | **[create-skill-file](./create-skill-file)** | Guides Claude in creating well-structured SKILL.md files with templates, examples, and best practices | - | `cp -r create-skill-file .claude/skills/` |
  11. | **[prompt-optimize](./prompt-optimize)** | Optimize your prompt with Claude | - | `cp -r prompt-optimize .claude/skills/` |
  12. | **[deep-reading-analyst](./deep-reading-analyst)** | Comprehensive framework for deep analysis using 10+ thinking models (SCQA, 5W2H, Critical Thinking, Mental Models, First Principles, etc.) | [🔗 GitHub](https://github.com/ginobefun/deep-reading-analyst-skill) | `cp -r deep-reading-analyst .claude/skills/` |
  13. | **[dry-refactoring](./dry-refactoring)** | Systematic code refactoring following DRY principle with 4-step workflow to eliminate code duplication | - | `cp -r dry-refactoring .claude/skills/` |
  14. | **[frontend-design](./frontend-design)** | Creates unique, production-grade frontend interfaces with exceptional design quality and creative aesthetics | - | `cp -r frontend-design .claude/skills/` |
  15. | **[mcp-builder](./mcp-builder)** | Guide for creating high-quality MCP servers that enable LLMs to interact with external services through tools | - | `cp -r mcp-builder .claude/skills/` |
  16. | **[daily-ai-news](./daily-ai-news)** | Aggregates and summarizes the latest AI news from multiple sources with concise briefs and direct links | - | `cp -r daily-ai-news .claude/skills/` |
  17. | **[fastgpt-workflow-generator](./fastgpt-workflow-generator)** | Generates production-ready FastGPT workflow JSON from natural language requirements with AI-powered template matching and three-layer validation | - | `cp -r fastgpt-workflow-generator .claude/skills/` |
  18. | **[planning-with-files](./planning-with-files)** | Manus-style workflow using persistent markdown files for planning, progress tracking, and knowledge storage with 3-file pattern | [🔗 GitHub](https://github.com/OthmanAdi/planning-with-files) | `cp -r planning-with-files .claude/skills/` |
  19. | **[local-diff-review](./local-diff-review)** | Pre-PR local code review skill based on git diff, applies the same three-dimension standards as pr-review (code quality, style guide, common issues checklist) | - | `cp -r local-diff-review .claude/skills/` |
  20. | **[software-copyright-writer](./software-copyright-writer)** | Prepares legitimate software copyright registration materials from real repositories, product pages, UI screenshots, source evidence, and module scopes | - | `cp -r software-copyright-writer .claude/skills/` |
  21. **Total:** 11 skills available
  22. ---
  23. ### Skill Details
  24. #### 🔧 create-skill-file
  25. **Versions:** [Chinese](./create-skill-file) / [English](./create-skill-file-EN)
  26. A meta-skill that teaches you how to create high-quality SKILL.md files for Claude.
  27. **What's included:**
  28. - ✅ Comprehensive writing guidelines
  29. - ✅ Ready-to-use templates (Basic & Workflow)
  30. - ✅ Real-world examples (Good & Bad practices)
  31. - ✅ Quality checklist and troubleshooting guide
  32. **Trigger Keywords:** `"create skill"`, `"write skill"`, `"SKILL.md"`, `"skill guidelines"`, `"best practices"`
  33. **Installation:**
  34. ```bash
  35. # Chinese version
  36. cp -r create-skill-file .claude/skills/
  37. # English version
  38. cp -r create-skill-file-EN .claude/skills/
  39. ```
  40. ---
  41. #### 🔧 prompt-optimize
  42. **Version:** [Chinese](./prompt-optimize)
  43. An expert prompt engineering skill that transforms Claude into "Alpha-Prompt" - a master prompt engineer who collaboratively crafts high-quality prompts through flexible dialogue.
  44. **What's included:**
  45. - ✅ Expert prompt engineering consultation
  46. - ✅ Advanced cognitive architectures (CoT, ToT, Self-Consistency, ReAct)
  47. - ✅ Security guardrails and safety considerations
  48. - ✅ Architecture upgrade suggestions for simple requirements
  49. - ✅ Collaborative dialogue-based prompt optimization
  50. **Key Features:**
  51. - **Flexible Communication**: Genuine two-way dialogue, not rigid templated questions
  52. - **Proactive Architecture Upgrades**: Suggests advanced techniques like Tree of Thought for creative tasks
  53. - **Security Awareness**: Provides safety recommendations for public-facing AI roles
  54. - **Quality Standards**: Delivers production-ready prompts with clear role definitions and structured outputs
  55. **Trigger Keywords:** `"optimize prompt"`, `"improve prompt"`, `"enhance AI instruction"`, `"prompt engineering"`, `"system instruction"`
  56. **Installation:**
  57. ```bash
  58. cp -r prompt-optimize .claude/skills/
  59. ```
  60. ---
  61. #### 🔧 deep-reading-analyst
  62. A professional skill that transforms surface-level reading into deep learning through systematic analysis using 10+ proven thinking frameworks.
  63. **For detailed documentation, see:** [Deep Reading Analyst Introduction](./skill-intro/depp-reading-analyst-intro.md)
  64. **🔗 GitHub Repository:** https://github.com/ginobefun/deep-reading-analyst-skill/tree/main
  65. **Installation:**
  66. ```bash
  67. cp -r deep-reading-analyst .claude/skills/
  68. ```
  69. ---
  70. #### 🔧 dry-refactoring
  71. **Version:** [Chinese](./dry-refactoring)
  72. A systematic skill that guides code refactoring following the DRY (Don't Repeat Yourself) principle through a proven 4-step workflow.
  73. **What's included:**
  74. - ✅ Step 1: Identify Repetition (obvious and semantic duplication)
  75. - ✅ Step 2: Abstract the Logic (functions, classes, constants)
  76. - ✅ Step 3: Replace Implementation (systematic replacement)
  77. - ✅ Step 4: Verify and Test (unit tests, integration tests, performance)
  78. **Key Features:**
  79. - **Comprehensive Coverage**: Handles copy-paste code, magic numbers, structural and logical repetition
  80. - **Step-by-Step Process**: Clear workflow from identification to verification
  81. - **Real-World Examples**: Complete e-commerce discount calculation refactoring case
  82. - **Best Practices**: Includes common pitfalls, testing strategies, and gradual refactoring approach
  83. **Trigger Keywords:** `"DRY"`, `"code duplication"`, `"refactor repetitive code"`, `"eliminate duplication"`, `"magic numbers"`, `"code smell"`, `"extract function"`
  84. **Installation:**
  85. ```bash
  86. cp -r dry-refactoring .claude/skills/
  87. ```
  88. ---
  89. #### 🔧 frontend-design
  90. **Version:** [Chinese](./frontend-design)
  91. An expert skill for creating unique, production-grade frontend interfaces with exceptional design quality that avoids generic AI aesthetics.
  92. **What's included:**
  93. - ✅ Design thinking framework (Purpose, Style, Constraints, Differentiation)
  94. - ✅ Typography guide (unique font selection, avoid common fonts)
  95. - ✅ Color & theme systems (CSS variables, accessibility)
  96. - ✅ Animation best practices (CSS animations, Framer Motion)
  97. - ✅ Spatial composition techniques (asymmetric layouts, overlapping elements)
  98. - ✅ Background & visual details (gradients, noise textures, glass morphism)
  99. **Key Features:**
  100. - **Bold Aesthetic Choices**: Guides selection of extreme, intentional design styles
  101. - **Anti-Generic AI Design**: Explicitly avoids overused fonts (Inter, Roboto) and cliché color schemes
  102. - **Implementation Complexity Matching**: Complex designs get complex code, minimal designs get precise code
  103. - **Production-Ready Code**: Semantic HTML, accessibility, responsive design, performance optimization
  104. **Trigger Keywords:** `"web component"`, `"landing page"`, `"dashboard"`, `"React component"`, `"UI design"`, `"beautify"`, `"frontend"`
  105. **Installation:**
  106. ```bash
  107. cp -r frontend-design .claude/skills/
  108. ```
  109. ---
  110. #### 🔧 mcp-builder
  111. **Version:** [English](./mcp-builder)
  112. A comprehensive guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools.
  113. **What's included:**
  114. - ✅ Phase 1: Deep Research and Planning (MCP design patterns, API coverage)
  115. - ✅ Phase 2: Tool Design (naming conventions, parameter schemas, error handling)
  116. - ✅ Phase 3: Implementation (FastMCP/Python, MCP SDK/TypeScript)
  117. - ✅ Phase 4: Testing and Publishing (integration tests, documentation)
  118. **Key Features:**
  119. - **Modern MCP Design Patterns**: Balances comprehensive API coverage with specialized workflow tools
  120. - **Framework-Specific Guidance**: Detailed examples for both FastMCP (Python) and MCP SDK (TypeScript)
  121. - **Quality Standards**: Emphasizes clear naming, actionable error messages, and proper context management
  122. - **Complete Workflow**: From initial research through testing and publishing
  123. **Supported Frameworks:**
  124. - FastMCP (Python) - Recommended for rapid development
  125. - MCP SDK (TypeScript/Node.js) - For JavaScript ecosystem integration
  126. **Trigger Keywords:** `"MCP server"`, `"Model Context Protocol"`, `"build MCP"`, `"FastMCP"`, `"MCP SDK"`, `"tool integration"`
  127. **Installation:**
  128. ```bash
  129. cp -r mcp-builder .claude/skills/
  130. ```
  131. ---
  132. #### 🔧 daily-ai-news
  133. **Version:** [Chinese & English](./daily-ai-news)
  134. An intelligent skill that aggregates and summarizes the latest AI news from multiple sources (AI news websites and web search) with concise briefs and direct links to original articles.
  135. **What's included:**
  136. - ✅ Phase 1: Information Gathering (WebSearch + mcp__web_reader__webReader)
  137. - ✅ Phase 2: Content Filtering (last 24-48 hours, deduplication)
  138. - ✅ Phase 3: Categorization (Major Announcements, Research, Industry, Tools, Policy)
  139. - ✅ Phase 4: Output Formatting (structured template with links)
  140. - ✅ Comprehensive news sources database (20+ sources)
  141. - ✅ Search query templates by category
  142. - ✅ Multiple output format templates
  143. **Key Features:**
  144. - **Multi-Source Aggregation**: Fetches from 3-5 major AI news sites plus web search
  145. - **Smart Filtering**: Keeps only recent news (24-48 hours) with major significance
  146. - **Structured Categorization**: 5 categories - Major Announcements, Research & Papers, Industry & Business, Tools & Applications, Policy & Ethics
  147. - **Direct Links**: Every news item includes a direct link to the original article
  148. - **Customization Options**: Focus areas, depth levels, time ranges, and format preferences
  149. - **Bilingual Support**: Works with both Chinese and English news sources
  150. **Trigger Keywords:** `"给我今天的AI资讯"`, `"today's AI news"`, `"AI updates"`, `"latest AI developments"`, `"daily AI briefing"`, `"AI industry news"`, `"artificial intelligence news"`
  151. **Installation:**
  152. ```bash
  153. cp -r daily-ai-news .claude/skills/
  154. ```
  155. ---
  156. #### 🔧 fastgpt-workflow-generator
  157. **Version:** [English](./fastgpt-workflow-generator)
  158. An intelligent skill that automatically generates production-ready FastGPT workflow JSON from natural language requirements using AI-powered semantic template matching and comprehensive three-layer validation.
  159. **What's included:**
  160. - ✅ Phase 1: Requirements Analysis (AI semantic extraction)
  161. - ✅ Phase 2: Template Matching (two-stage: coarse + fine filtering)
  162. - ✅ Phase 3: JSON Generation (NodeId generation, auto-layout, reference handling)
  163. - ✅ Phase 4: Validation (format, connections, logic completeness)
  164. - ✅ Phase 5: Incremental Modification (add/remove/modify nodes)
  165. - ✅ Built-in templates (document translation, sales training, resume screening, financial news)
  166. - ✅ Complete references (40+ node types, validation rules, JSON structure specs)
  167. - ✅ Validation script (Node.js)
  168. **Key Features:**
  169. - **AI-Powered Template Matching**: Two-stage semantic matching (metadata + AI analysis) to find the most similar template
  170. - **Automatic JSON Generation**: Generates complete workflow with semantic NodeIds, auto-layout positions, and proper references
  171. - **Three-Layer Validation**: Format (JSON structure) → Connections (node references) → Logic (workflow completeness)
  172. - **Incremental Modification Support**: Add, delete, or modify nodes in existing workflows
  173. - **Built-in Templates**: 4 production-ready templates covering different domains and complexities
  174. - **Portable Design**: Self-contained with relative paths, works in any project
  175. **Built-in Templates:**
  176. - `文档翻译助手.json` - Simple workflow (document processing)
  177. - `销售陪练大师.json` - Medium complexity (conversational AI)
  178. - `简历筛选助手_飞书.json` - Complex workflow (data + external integration)
  179. - `AI金融日报.json` - Scheduled trigger + multi-agent (news aggregation)
  180. **Reference Format:**
  181. - Array reference: `["nodeId", "key"]` - for direct values
  182. - Template reference: `{{$nodeId.key$}}` - for string interpolation (double braces!)
  183. **Trigger Keywords:** `"create FastGPT workflow"`, `"generate workflow JSON"`, `"design FastGPT application"`, `"工作流"`, `"workflow automation"`, `"multi-agent systems"`, `"FastGPT templates"`
  184. **Installation:**
  185. ```bash
  186. cp -r fastgpt-workflow-generator .claude/skills/
  187. ```
  188. **Quick Start:**
  189. ```
  190. # Example 1: Create from natural language
  191. "生成一个旅游规划的工作流,收集用户的目的地、天数、预算等信息,然后AI生成详细的旅游规划"
  192. # Example 2: Modify existing workflow
  193. "在现有问答工作流前添加知识库搜索节点"
  194. # Example 3: Validate workflow
  195. "验证这个workflow JSON是否正确"
  196. ```
  197. ---
  198. #### 🔧 planning-with-files
  199. **Version:** [English](./planning-with-files)
  200. A Manus-style workflow skill that transforms how you work with Claude by using persistent markdown files as "working memory on disk" for planning, progress tracking, and knowledge storage.
  201. **🔗 Reference:** Based on [OthmanAdi/planning-with-files](https://github.com/OthmanAdi/planning-with-files)
  202. **What's included:**
  203. - ✅ 3-File Pattern System (task_plan.md, notes.md, deliverable.md)
  204. - ✅ Task plan template with phases, decisions, and errors tracking
  205. - ✅ Notes template for research and findings
  206. - ✅ Progress tracking with checkboxes and status updates
  207. - ✅ Knowledge persistence outside attention window
  208. - ✅ Complete workflow loop patterns
  209. - ✅ Real-world examples and anti-patterns
  210. **Key Features:**
  211. - **Persistent Memory**: Store plans, research, and decisions in files instead of context window
  212. - **3-File Pattern**: Systematic approach with task_plan.md (progress), notes.md (research), and deliverable.md (output)
  213. - **Structured Tracking**: Phase-based planning with checkboxes, decisions log, and error tracking
  214. - **Loop-Based Workflow**: Read → Work → Update cycle keeps goals fresh in attention
  215. - **Context Management**: Prevents context stuffing by storing information in files
  216. - **Multi-Session Support**: Persist work across conversation sessions
  217. **The 3-File Pattern:**
  218. - **task_plan.md** - Track phases and progress (update after each phase)
  219. - **notes.md** - Store findings and research (during research)
  220. - **[deliverable].md** - Final output (at completion)
  221. **Core Workflow Loop:**
  222. ```
  223. 1. Create task_plan.md with goal and phases
  224. 2. Research → save to notes.md → update task_plan.md
  225. 3. Read notes.md → create deliverable → update task_plan.md
  226. 4. Deliver final output
  227. ```
  228. **When to Use:**
  229. - Starting complex multi-step projects
  230. - Research tasks requiring information gathering
  231. - Tasks requiring progress tracking across sessions
  232. - User mentions "planning", "organizing work", "tracking progress"
  233. - Need structured output with clear phases
  234. **Trigger Keywords:** `"complex task"`, `"multi-step project"`, `"planning"`, `"organize work"`, `"track progress"`, `"research task"`, `"structured output"`
  235. **Installation:**
  236. ```bash
  237. cp -r planning-with-files .claude/skills/
  238. ```
  239. **Quick Start:**
  240. ```
  241. # Example: Plan a feature implementation
  242. "I need to implement a new authentication system"
  243. → Claude creates:
  244. - task_plan.md (phases: design, implementation, testing, deployment)
  245. - notes.md (stores API research and design decisions)
  246. - Updates plan after each phase completion
  247. ```
  248. ---
  249. #### 🔧 local-diff-review
  250. **Version:** [Chinese](./local-diff-review)
  251. A pre-PR self-review skill that runs git diff locally and applies the same three-dimension review standards as a formal PR review — no custom weakening of criteria allowed.
  252. **What's included:**
  253. - ✅ Code quality standards (`code-quality-standards.md`)
  254. - ✅ Common issues checklist (`common-issues-checklist.md`)
  255. - ✅ 5 diff scope modes (unstaged, staged, HEAD, branch-vs-baseline, specific commit)
  256. - ✅ Three-level issue grading (🔴 Critical / 🟡 Suggested / 🟢 Optional)
  257. - ✅ Structured output: change overview → issue list → overall conclusion
  258. **Key Features:**
  259. - **Review First, Fix Later**: Outputs an issue list by default; only modifies code when explicitly asked
  260. - **Actionable Results**: Every issue includes file path, line number, risk explanation, and a concrete fix suggestion
  261. - **Regression Risk Summary**: Highlights i18n, type safety, edge cases, compatibility, and performance impact
  262. - **Consistent Standards**: Strictly reuses the same criteria as `pr-review`; no softened rules
  263. **Trigger Keywords:** `"review 我本地改动"`, `"local diff review"`, `"review 我当前改动"`, `"git diff review"`, `"暂存区 review"`, `"分支差异审查"`, `"提 PR 前 review"`
  264. **Installation:**
  265. ```bash
  266. cp -r local-diff-review .claude/skills/
  267. ```
  268. **Quick Start:**
  269. ```
  270. # Example 1: Review all unstaged changes
  271. "请按 PR 标准 review 我本地当前改动。"
  272. # Example 2: Review only staged changes
  273. "请 review 我的 git diff --staged,只给问题清单。"
  274. # Example 3: Review branch diff against baseline
  275. "请按 FastGPT 规范审查我当前分支相对 upstream/v4.14.7-dev 的改动。"
  276. ```
  277. ---
  278. #### 🔧 software-copyright-writer
  279. **Version:** [Chinese](./software-copyright-writer)
  280. A material-preparation skill for legitimate software copyright registration. It analyzes real repositories, product pages, running UI evidence, module scopes, and source ranges to propose reasonable registration topics, generate long-form Chinese application content, assemble source-code excerpts, capture web evidence, and export `.docx` deliverables.
  281. **Compliance Boundary:**
  282. - Only use it for real software projects with authentic source code and actual running interfaces
  283. - Do not use it to fabricate software, screenshots, ownership, domain filings, trademarks, or features
  284. - If evidence is missing or the requested topic count is unrealistic, the skill must report gaps and risks instead of inventing material
  285. **What's included:**
  286. - ✅ Topic splitting rules for coarse/fine software copyright applications
  287. - ✅ Intake checklist for owner name, Logo, copyright notice, ICP/domain evidence, screenshots, and source-code constraints
  288. - ✅ Standard and extended Chinese content templates targeting 30-page style application materials
  289. - ✅ Source-code assembly script using the front/back 1500-line rule
  290. - ✅ Playwright-based web evidence screenshot script with login-state support
  291. - ✅ Markdown-to-DOCX export script for content and source documents
  292. **Requirements for helper scripts:**
  293. - Python 3 with `python-docx` for `.docx` export (`python3 -m pip install -r scripts/requirements.txt`)
  294. - Node.js, Playwright, and Chrome/Chromium for screenshot evidence capture (`npm install --prefix scripts`)
  295. - Optional environment variables: `SOFTWARE_COPYRIGHT_WRITER_PYTHON`, `SOFTWARE_COPYRIGHT_WRITER_NODE`, `SOFTWARE_COPYRIGHT_WRITER_NODE_MODULES`, `SOFTWARE_COPYRIGHT_WRITER_CHROME_PATH`
  296. **Trigger Keywords:** `"软著"`, `"软件著作权"`, `"生成软著材料"`, `"软著正文"`, `"源码原文"`, `"软件著作权登记"`, `"copyright registration material"`
  297. **Installation:**
  298. ```bash
  299. cp -r software-copyright-writer .claude/skills/
  300. ```
  301. **Quick Start:**
  302. ```
  303. # Example 1: Evaluate whether a repository can support multiple registration topics
  304. "请根据这个仓库和官网,判断能拆几份软著,并给出主题边界和截图清单。"
  305. # Example 2: Generate full materials after evidence is ready
  306. "按已确认的主题生成软著正文、源码原文,并导出 docx。"
  307. ```
  308. ---
  309. ## 🚀 How to Use
  310. ### Installation Steps
  311. 1. **Navigate to your project**
  312. ```bash
  313. cd /path/to/your/project
  314. ```
  315. 2. **Create the skills directory** (if not exists)
  316. ```bash
  317. mkdir -p .claude/skills
  318. ```
  319. 3. **Copy the skill you want**
  320. Example - Adding the skill creation guide:
  321. ```bash
  322. # clone this repository if you haven't already
  323. git clone https://github.com/YYH211/Claude-meta-skill.git
  324. # For Chinese version
  325. cp -r /Claude-meta-skill/create-skill-file .claude/skills/
  326. # OR for English version
  327. cp -r /Claude-meta-skill/create-skill-file-EN .claude/skills/
  328. ```
  329. 4. **Verify installation**
  330. ```bash
  331. ls .claude/skills/
  332. ```
  333. You should see your copied skill directory:
  334. ```
  335. .claude/skills/
  336. ├── create-skill-file/ (or create-skill-file-EN/)
  337. │ ├── SKILL.md
  338. │ ├── templates/
  339. │ └── examples/
  340. ```
  341. ### Verifying Configuration
  342. #### ✅ Method 1: Check File Structure
  343. ```bash
  344. cat .claude/skills/create-skill-file/SKILL.md | head -10
  345. ```
  346. Expected output should show the YAML frontmatter:
  347. ```yaml
  348. ---
  349. name: create-skill-file
  350. description: Guides Claude in creating well-structured SKILL.md files...
  351. ---
  352. ```
  353. #### ✅ Method 2: Test with Claude
  354. Start Claude Code and try triggering the skill:
  355. ```bash
  356. claude
  357. ```
  358. Then ask:
  359. - "Help me create a new skill"
  360. - "I want to write a SKILL.md file"
  361. - "Guide me on skill best practices"
  362. If Claude provides structured guidance referencing templates and examples, the skill is working! 🎉
  363. #### ✅ Method 3: List All Skills
  364. You can check what skills are available by asking Claude:
  365. ```
  366. What skills do you have access to?
  367. ```
  368. claude response:
  369. ```
  370. 1. create-skill-file (Skill 文件创建指南)
  371. 功能: 指导创建结构良好的 SKILL.md 文件
  372. 触发场景:
  373. - 当你询问如何创建 Skill
  374. - 需要编写 SKILL.md 文档
  375. - 想了解 Skill 编写最佳实践
  376. 核心能力:
  377. - 提供清晰的命名、结构和内容组织指南
  378. - 包含模板和示例
  379. - 质量检查清单
  380. - 常见问题解答
  381. 位置: .claude/skill/create-skill-file/SKILL.md
  382. ```
  383. Or manually check:
  384. ```bash
  385. find .claude/skills -name "SKILL.md" -exec grep -A 2 "^name:" {} \;
  386. ```
  387. ---
  388. ## 💡 Usage Examples
  389. ### Example 1: Creating a New Skill
  390. ```
  391. You: "I need to create a skill for managing Docker deployments"
  392. Claude: [Uses create-skill-file skill]
  393. "I'll help you create a Docker deployment skill. Let's start by following
  394. the skill creation guidelines..."
  395. ```
  396. ### Example 2: Improving Existing Skills
  397. ```
  398. You: "Can you review my database-migration skill and suggest improvements?"
  399. Claude: [References best practices from create-skill-file]
  400. "Let me review your skill against the best practices checklist..."
  401. ```
  402. ---
  403. ## 🛠️ Troubleshooting
  404. ### Skill Not Working
  405. | Problem | Solution |
  406. |---------|----------|
  407. | Claude doesn't use the skill | Verify the skill is in `.claude/skills/` directory |
  408. | Skill file not found | Check that SKILL.md exists and has valid frontmatter |
  409. | Wrong skill activates | Make description more specific with unique keywords |
  410. | Skill seems ignored | Try using explicit trigger keywords from the description |
  411. ### Common Issues
  412. **Q: Do I need to restart Claude after adding a skill?**
  413. A: Typically yes. Restart your Claude Code session after adding new skills.
  414. **Q: Can I modify the skills?**
  415. A: Absolutely! Feel free to customize them for your project needs.
  416. **Q: How many skills can I have?**
  417. A: No hard limit, but keep it manageable (5-10 skills per project is typical).
  418. **Q: Can I use multiple skills together?**
  419. A: Yes! Claude can use multiple skills in the same conversation as needed.
  420. ---
  421. ## 🚧 Coming Soon
  422. This repository will grow with more useful skills:
  423. - [ ] **code-review-workflow** - Systematic code review process
  424. - [ ] **api-documentation-generator** - Generate API docs from code
  425. - [ ] **testing-strategy** - Guide for writing comprehensive tests
  426. - [ ] **deployment-checklist** - Pre-deployment verification
  427. - [ ] **refactoring-guide** - Code refactoring best practices
  428. **Want to contribute?** Create a skill following the guidelines in `create-skill-file` and submit it!
  429. ---
  430. ## 📚 Resources
  431. - [Claude Code Documentation](https://docs.claude.com/en/docs/claude-code)
  432. - [Agent Skills Official Guide](https://docs.claude.com/en/docs/agents-and-tools/agent-skills)
  433. - [Best Practices](https://docs.claude.com/en/docs/agents-and-tools/agent-skills/best-practices)
  434. ---
  435. ## ⭐ Star History
  436. [![Star History Chart](https://api.star-history.com/svg?repos=YYH211/Claude-meta-skill&type=Date)](https://star-history.com/#YYH211/Claude-meta-skill&Date)
  437. ---
  438. ## 🤝 Partner Projects
  439. ### FastGPT
  440. [FastGPT](https://github.com/labring/FastGPT) is a knowledge-based platform built on the LLM, offers out-of-the-box data processing and model invocation capabilities, allows for workflow orchestration through Flow visualization.
  441. ---
  442. ## 🤝 Contributing
  443. Have a useful skill to share? We'd love to include it!
  444. 1. Create your skill following the `create-skill-file` guidelines
  445. 2. Test it thoroughly in real projects
  446. 3. Include clear documentation and examples
  447. 4. Submit via pull request or issue
  448. ---
  449. ## 📄 License
  450. Free to use for any purpose. Customize and adapt as needed for your projects.
  451. ---
  452. **Happy coding with Claude!** 🚀

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.