claude-code-config-LANG

# Claude Code Configuration > Based on the actual configuration of Claude Code on macOS, with desensitization. Applicable to Claude Code CLI / Desktop / IDE extensions. ## 📁 Directory Structure ``` claude-code-config/ ├── README.md # This document ├── RTK.md # RTK (Rust Token Killer) configuration ├── settings.json # Main configuration (desensitized) ├── agents/ # Custom sub-agents (10) │ ├── bug-analyzer.md │ ├── code-explore.md │ ├── code-reviewer.md │ ├── dev-planner.md │ ├── general-purpose.md │ ├── impl-plan.md │ ├── story-generator.md │ ├── tech-lead.md │ ├── test-writer.md │ └── ui-sketcher.md ├── commands/ # Custom commands (5) │ ├── clean-code-mentor.md │ ├── commit.md │ ├── modern-dev.md │ ├── modern-ui-builder.md │ └── syncthing.md ├── mcp-servers/ # MCP server scripts │ └── qwen_vl_mcp_server.py ├── output-styles/ # Output styles (2) │ ├── coding-vibes.md │ └── structural-thinking.md ├── plugins/ # Plugin configurations │ ├── installed_plugins.json │ └── known_marketplaces.json └── skills/ # Skills (41, complete content) ├── humanizer/ # AI-text refinement ├── web-access/ # Web access/CDP browser ├── docx/ # Word document operations ├── pdf/ # PDF operations ├── pptx/ # PowerPoint operations ├── xlsx/ # Excel operations ├── lark-*/ # Lark suite (24) └── ... # More skills ``` *** ## 🔧 Installation Steps ### 1. Basic Configuration ```bash # Copy the configuration file to the Claude Code configuration directory cp settings.json ~/.claude/settings.json # Edit settings.json, replace all YOUR_* placeholders # - YOUR_ANTHROPIC_AUTH_TOKEN: Your API key # - YOUR_ANTHROPIC_BASE_URL: API endpoint URL # - YOUR_*_MODEL: Model ID ``` ### 2. Sub-agents (Agents) ```bash cp agents/*.md ~/.claude/agents/ ``` ### 3. Custom Commands (Commands) ```bash cp commands/*.md ~/.claude/commands/ ``` ### 4. Output Styles (Output Styles) ```bash cp output-styles/*.md ~/.claude/output-styles/ ``` ### 5. Plugins (Plugins) ```bash cp plugins/*.json ~/.claude/plugins/ # Then install plugins through Claude Code: claude install pua@pua-skills claude install skill-creator@claude-plugins-official claude install superpowers@claude-plugins-official claude install vercel-plugin@vercel-vercel-plugin claude install clangd-lsp@claude-plugins-official claude install claude-mem@thedotmack ``` ### 6. MCP Server ```bash mkdir -p ~/mcp_servers cp mcp-servers/qwen_vl_mcp_server.py ~/mcp_servers/ # Edit the script, replace YOUR_SERVER_IP with the actual address ``` ### 7. Skills (Skills) ```bash # Method 1: Symbolic link (recommended, easy to update) for skill in skills/*/; do ln -sf "$(pwd)/$skill" ~/.claude/skills/ done # Method 2: Copy cp -r skills/* ~/.claude/skills/ # Method 3: Install as needed ln -sf "$(pwd)/skills/humanizer" ~/.claude/skills/ ln -sf "$(pwd)/skills/web-access" ~/.claude/skills/ ``` ### 8. RTK (Rust Token Killer) ```bash cp RTK.md ~/.claude/RTK.md # Install rtk: cargo install rtk or download from GitHub ``` *** ## 📋 Complete Configuration List ### 🤖 Sub-agents (10) | **Agent** | **Purpose** | **Tool permissions** | | ----------------- | ----------------------------------- | ------------------------------------------------------------ | | `bug-analyzer` | Deep bug analysis, tracking variable states, locating root causes | Glob, Grep, Read, Bash | | `code-explore` | Quick code library exploration, supporting quick/medium/thorough depth | Glob, Grep, Read, WebFetch, WebSearch, Bash | | `code-reviewer` | Professional code review: quality, security, performance, best practices | Glob, Grep, Read, Bash | | `dev-planner` | Development planning: task decomposition, dependency analysis, time estimation | Default tools | | `general-purpose` | General tasks: complex problem research, multi-step task execution | Glob, Grep, Read, Write, Edit, Bash, WebFetch, WebSearch | | `impl-plan` | Architecture planning: implementation plan design, architecture trade-offs | Glob, Grep, Read, WebFetch, WebSearch, Bash | | `story-generator` | User story generation: generating stories from diff/PRD/requirements | Bash, Glob, Grep, Read, WebFetch, TodoWrite, WebSearch | | `tech-lead` | Technical lead: requirement decomposition, task definition, multi-agent coordination | Glob, Grep, Read, WebFetch, WebSearch | | `test-writer` | Test writing: unit/integration/end-to-end testing | Glob, Grep, Read, Write, Edit, Bash | | `ui-sketcher` | UI blueprint engineer: ASCII interface design, interaction specifications | Bash, Glob, Grep, Read, WebFetch, TodoWrite, WebSearch + MCP | ### ⚡ Custom Commands (5) | **Command** | **Purpose** | | ------------------- | ------------------------------------------ | | `clean-code-mentor` | Programming mentor: production-level code teaching, line-by-line explanation | | `commit` | Conventional Commits规范提交，支持emoji前缀 | | `modern-dev` | Enforced modern technology stack (React 19, Rust Axum, Bun, etc.) | | `modern-ui-builder` | Production-level UI component construction (React/Vue + Tailwind + Shadcn) | | `syncthing` | Syncthing file synchronization expert (CLI/REST API/configuration) | ### 🔌 Plugins (6) | **Plugin** | **Source** | **Version** | **Description** | **GitHub** | | --------------- | ----------------------- | ------ | ------------ | --------------------------------------------------------------------------------------------- | | `pua` | pua-skills | 2.3.0 | PUA prompt optimization | ​[tanweai/pua](https://github.com/tanweai/pua)​ | | `skill-creator` | claude-plugins-official | - | Skill creator | ​[anthropics/claude-plugins-official](https://github.com/anthropics/claude-plugins-official)​ | | `superpowers` | claude-plugins-official | 6.0.3 | Enhanced capabilities | ​[anthropics/claude-plugins-official](https://github.com/anthropics/claude-plugins-official)​ | | `vercel-plugin` | vercel-vercel-plugin | 0.25.0 | Vercel deployment integration | ​[vercel/vercel-plugin](https://github.com/vercel/vercel-plugin)​ | | `clangd-lsp` | claude-plugins-official | 1.0.0 | C/C++ LSP support | ​[anthropics/claude-plugins-official](https://github.com/anthropics/claude-plugins-official)​ | | `claude-mem` | thedotmack | 13.8.1 | Memory management system | ​[thedotmack/claude-mem](https://github.com/thedotmack/claude-mem)​ | ### 🛠️ Skills (41) #### Native Installation (1) | **Skill** | **Description** | **GitHub** | | ----------- | --------------------------- | --------------------------------------------------------- | | `humanizer` | AI-text refinement, detecting 33 AI writing modes | ​[blader/humanizer](https://github.com/blader/humanizer)​ | #### cc-switch Skill Package (15) > From `~/.cc-switch/skills/`, installed via symbolic links | **Skill** | **Description** | **GitHub** | | ------------------------------ | --------------- | ----------------------------------------------------------- | | `clean-code-mentor` | Production-level code teaching mentor | - | | `docx` | Word document read/write/edit | - | | `excalidraw-diagram-generator` | Excalidraw diagram generation | - | | `frontend-design` | High-quality frontend interface design | - | | `git-commit` | Intelligent Git commit | - | | `ima-skill` | Tencent IMA knowledge base operation | - | | `modern-dev` | Enforced modern technology stack | - | | `modern-ui-builder` | Modern UI component construction | - | | `pdf` | Full-featured PDF operation | - | | `pptx` | PowerPoint read/write/edit | - | | `prd` | Product requirement document generation | - | | `refactor` | Surgical code refactoring | - | | `skill-creator` | Skill creation and optimization | - | | `web-access` | Web access/crawler/CDP browser | ​[eze-is/web-access](https://github.com/eze-is/web-access)​ | | `xlsx` | Excel spreadsheet operation | - | #### Lark Skills (24) > From `~/.agents/skills/`, requires `lark-cli` binary dependency | **Skill** | **Description** | | ------------------------------- | --------------- | | `lark-shared` | Shared foundation: authentication, authorization, identity switching | | `lark-approval` | Approval process management | | `lark-attendance` | Attendance query | | `lark-base` | Multi-dimensional table operation | | `lark-calendar` | Calendar and meeting management | | `lark-contact` | Contact query | | `lark-doc` | Cloud document creation and editing | | `lark-drive` | Cloud disk file management | | `lark-event` | Real-time event listening | | `lark-im` | Instant messaging | | `lark-mail` | Email management | | `lark-markdown` | Markdown file operation | | `lark-minutes` | Meeting minutes/recording | | `lark-okr` | OKR goal management | | `lark-openapi-explorer` | OpenAPI exploration | | `lark-sheets` | Spreadsheet operation | | `lark-skill-maker` | Custom skill creation | | `lark-slides` | Slide creation and editing | | `lark-task` | Task management | | `lark-vc` | Video conference query | | `lark-whiteboard` | Whiteboard operation | | `lark-wiki` | Knowledge base management | | `lark-workflow-meeting-summary` | Meeting summary aggregation | | `lark-workflow-standup-report` | Stand-up report generation | ### 🌐 MCP Server (2) | **Server** | **Description** | **Dependency** | | --------- | --------------------- | --------------------------- | | `qwen-vl` | Qwen3.5-9B-VL multimodal visual model | Locally deployed Qwen service | | `stitch` | Stitch design to code | npx @\_davideast/stitch-mcp | ### 📝 Output Style (2) | **Style** | **Description** | | --------------------- | ------------------- | | `coding-vibes` | Energetic coding companion, supports multiple languages | | `structural-thinking` | Structured thinking, clear expression | ### 📐 Rules (20+) Core rules include: * Communication and comments in English * Modern tech stack enforcement (React 19, Rust Axum, Bun, etc.) * Type safety (strict mode, no any) * Asynchronous priority (all I/O must be asynchronous) * Production-level code standards (no toy code) * Programming mentor role (line-by-line explanation, analogy) ### 🔗 External Dependencies | **Dependency** | **GitHub** | **Description** | | ----------------------- | --------------------------------------------------------------------------------------------- | -------------------------- | | RTK | - | Rust Token Killer, CLI proxy optimization | | humanizer | ​[blader/humanizer](https://github.com/blader/humanizer)​ | AI-free text refinement | | web-access | ​[eze-is/web-access](https://github.com/eze-is/web-access)​ | Web access/CDP browser | | claude-plugins-official | ​[anthropics/claude-plugins-official](https://github.com/anthropics/claude-plugins-official)​ | Official plugin market | | anthropic-agent-skills | ​[anthropics/skills](https://github.com/anthropics/skills)​ | Official skill library | | pua | ​[tanweai/pua](https://github.com/tanweai/pua)​ | PUA prompt optimization | | vercel-plugin | ​[vercel/vercel-plugin](https://github.com/vercel/vercel-plugin)​ | Vercel integration | | claude-mem | ​[thedotmack/claude-mem](https://github.com/thedotmack/claude-mem)​ | Memory management system | | lark-cli | - | Lark CLI tool (24 skill dependencies) | *** ## ⚠️ Desensitization Description The following content has been replaced with placeholders: | **Original Content** | **Placeholder** | | ------------------- | --------------------------- | | Anthropic API Token | `YOUR_ANTHROPIC_AUTH_TOKEN` | | API Base URL | `YOUR_ANTHROPIC_BASE_URL` | | Model ID | `YOUR_*_MODEL` | | Syncthing API Key | `YOUR_SYNCTHING_API_KEY` | | Syncthing Device ID | `YOUR_DEVICE_ID` | | Intranet IP Address | `YOUR_SERVER_IP` | | ZeroTier IP | `YOUR_ZEROTIER_IP` | | settings.local.json | Not included (contains historical permission records) | | credentials.json | Not included (authentication credentials) | | history.jsonl | Not included (session history) | *** ## 📖 Usage Suggestions 1. **Progressive Installation**: Install basic configuration and agents first, then add plugins and skills 2. **Hermes Agent**: Skill configuration can be migrated to Hermes, and Hermes has Lark message channel, but ClaudeCode is still unmatched in coding tasks 3. **Custom Rules**: Adjust the `rules` array in `settings.json` according to team habits 4. **MCP Server**: `qwen-vl` requires local deployment of Qwen model, `stitch` can be used directly 5. **Lark Skills**: Need to install `lark-cli` and complete authentication 6. **RTK**: Optional installation, used to reduce token consumption for CLI operations 7. **PUA**: Manually call when task execution results are not ideal 8. **Superpowers**: Manually call superpowers:brainstorming to assist in clarifying requirements during project construction 9. **Humanizer**: Used for text tasks to remove AI flavor