[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-lst97--claude-code-sub-agents":3,"tool-lst97--claude-code-sub-agents":64},[4,17,27,35,48,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",138956,2,"2026-04-05T11:33:21",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,43,44,45,15,46,26,13,47],"数据工具","视频","插件","其他","音频",{"id":49,"name":50,"github_repo":51,"description_zh":52,"stars":53,"difficulty_score":10,"last_commit_at":54,"category_tags":55,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,46],{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},2181,"OpenHands","OpenHands\u002FOpenHands","OpenHands 是一个专注于 AI 驱动开发的开源平台，旨在让智能体（Agent）像人类开发者一样理解、编写和调试代码。它解决了传统编程中重复性劳动多、环境配置复杂以及人机协作效率低等痛点，通过自动化流程显著提升开发速度。\n\n无论是希望提升编码效率的软件工程师、探索智能体技术的研究人员，还是需要快速原型验证的技术团队，都能从中受益。OpenHands 提供了灵活多样的使用方式：既可以通过命令行（CLI）或本地图形界面在个人电脑上轻松上手，体验类似 Devin 的流畅交互；也能利用其强大的 Python SDK 自定义智能体逻辑，甚至在云端大规模部署上千个智能体并行工作。\n\n其核心技术亮点在于模块化的软件智能体 SDK，这不仅构成了平台的引擎，还支持高度可组合的开发模式。此外，OpenHands 在 SWE-bench 基准测试中取得了 77.6% 的优异成绩，证明了其解决真实世界软件工程问题的能力。平台还具备完善的企业级功能，支持与 Slack、Jira 等工具集成，并提供细粒度的权限管理，适合从个人开发者到大型企业的各类用户场景。",70612,"2026-04-05T11:12:22",[26,15,13,45],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":70,"readme_en":71,"readme_zh":72,"quickstart_zh":73,"use_case_zh":74,"hero_image_url":75,"owner_login":76,"owner_name":77,"owner_avatar_url":78,"owner_bio":79,"owner_company":80,"owner_location":81,"owner_email":82,"owner_twitter":80,"owner_website":83,"owner_url":84,"languages":80,"stars":85,"forks":86,"last_commit_at":87,"license":88,"difficulty_score":23,"env_os":89,"env_gpu":90,"env_ram":90,"env_deps":91,"category_tags":102,"github_topics":103,"view_count":23,"oss_zip_url":80,"oss_zip_packed_at":80,"status":16,"created_at":110,"updated_at":111,"faqs":112,"releases":133},2521,"lst97\u002Fclaude-code-sub-agents","claude-code-sub-agents","Collection of specialized AI subagents for Claude Code for personal use (full-stack development).","claude-code-sub-agents 是一套专为 Claude Code 设计的开源子智能体集合，旨在通过引入 33 个具备特定领域专长的 AI 助手，全面增强全栈开发的工作流。它并非一个独立的软件，而是作为 Claude Code 的功能扩展插件存在，能够根据任务上下文自动调度或响应用户调用，提供从前端界面构建到后端架构设计的专业支持。\n\n这套工具主要解决了通用 AI 助手在应对复杂、垂直技术领域时可能出现的“博而不精”问题。通过将大任务拆解并委派给精通 React、Next.js、Python、Go 或云基础设施等特定技术的专家型子智能体，它显著提升了代码生成的准确性、规范性和执行效率。无论是处理响应式布局、优化数据库模式，还是配置 CI\u002FCD 流水线及排查生产环境故障，claude-code-sub-agents 都能提供更具针对性的解决方案，并内置了代码审查与验证机制以保障交付质量。\n\n该工具特别适合从事全栈开发、DevOps 运维以及系统架构设计的专业开发者使用。如果你希望在使用 Claude Code 进行日常编码、遗留系统现代化改造或云资源管理时获得更精准的智能辅","claude-code-sub-agents 是一套专为 Claude Code 设计的开源子智能体集合，旨在通过引入 33 个具备特定领域专长的 AI 助手，全面增强全栈开发的工作流。它并非一个独立的软件，而是作为 Claude Code 的功能扩展插件存在，能够根据任务上下文自动调度或响应用户调用，提供从前端界面构建到后端架构设计的专业支持。\n\n这套工具主要解决了通用 AI 助手在应对复杂、垂直技术领域时可能出现的“博而不精”问题。通过将大任务拆解并委派给精通 React、Next.js、Python、Go 或云基础设施等特定技术的专家型子智能体，它显著提升了代码生成的准确性、规范性和执行效率。无论是处理响应式布局、优化数据库模式，还是配置 CI\u002FCD 流水线及排查生产环境故障，claude-code-sub-agents 都能提供更具针对性的解决方案，并内置了代码审查与验证机制以保障交付质量。\n\n该工具特别适合从事全栈开发、DevOps 运维以及系统架构设计的专业开发者使用。如果你希望在使用 Claude Code 进行日常编码、遗留系统现代化改造或云资源管理时获得更精准的智能辅助，它将是一个得力的助手。其核心亮点在于智能化的多智能体编排能力，能够无缝协调不同领域的专家协同工作，既优化了资源利用，又确保了从 UI 设计到云端部署的全链路开发体验更加流畅高效。","# Claude Code Subagents Collection\n\nA comprehensive collection of 33 specialized AI subagents for [Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code), designed to enhance development workflows with domain-specific expertise and intelligent automation.\n\n## 🚀 Overview\n\nThis repository contains a curated set of specialized subagents that extend Claude Code's capabilities across the entire software development lifecycle. Each subagent is an expert in a specific domain, automatically invoked based on context analysis or explicitly called when specialized expertise is needed.\n\n### Key Features\n\n- **🤖 Intelligent Auto-Delegation**: Claude Code automatically selects optimal agents based on task context\n- **🔧 Domain Expertise**: Each agent specializes in specific technologies, patterns, and best practices\n- **🔄 Multi-Agent Orchestration**: Seamless coordination between agents for complex workflows\n- **📊 Quality Assurance**: Built-in review and validation patterns across all domains\n- **⚡ Performance Optimized**: Agents designed for efficient task completion and resource utilization\n\n## Available Subagents\n\nAgents are now organized into logical categories for easier navigation:\n\n### 🏗️ [Development](agents\u002Fdevelopment\u002F)\n\n**Frontend & UI Specialists**\n\n- **[frontend-developer](agents\u002Fdevelopment\u002Ffrontend-developer.md)** - Build React components, implement responsive layouts, and handle client-side state management\n- **[ui-designer](agents\u002Fdevelopment\u002Fui-designer.md)** - Creative UI design focused on user-friendly interfaces\n- **[ux-designer](agents\u002Fdevelopment\u002Fux-designer.md)** - User experience design and interaction optimization\n- **[react-pro](agents\u002Fdevelopment\u002Freact-pro.md)** - Expert React development with hooks, performance optimization, and best practices\n- **[nextjs-pro](agents\u002Fdevelopment\u002Fnextjs-pro.md)** - Next.js specialist for SSR, SSG, and full-stack React applications\n\n**Backend & Architecture**\n\n- **[backend-architect](agents\u002Fdevelopment\u002Fbackend-architect.md)** - Design RESTful APIs, microservice boundaries, and database schemas\n- **[full-stack-developer](agents\u002Fdevelopment\u002Ffull-stack-developer.md)** - End-to-end web application development from UI to database with seamless integration\n\n**Language Specialists**\n\n- **[python-pro](agents\u002Fdevelopment\u002Fpython-pro.md)** - Write idiomatic Python code with advanced features and optimizations\n- **[golang-pro](agents\u002Fdevelopment\u002Fgolang-pro.md)** - Write idiomatic Go code with goroutines, channels, and interfaces\n- **[typescript-pro](agents\u002Fdevelopment\u002Ftypescript-pro.md)** - Advanced TypeScript development with type safety and modern patterns\n\n**Platform & Mobile**\n\n- **[mobile-developer](agents\u002Fdevelopment\u002Fmobile-developer.md)** - Develop React Native or Flutter apps with native integrations\n- **[electron-pro](agents\u002Fdevelopment\u002Felectorn-pro.md)** - Electron desktop application development and cross-platform solutions\n\n**Developer Experience**\n\n- **[dx-optimizer](agents\u002Fdevelopment\u002Fdx-optimizer.md)** - Developer Experience specialist that improves tooling, setup, and workflows\n- **[legacy-modernizer](agents\u002Fdevelopment\u002Flegacy-modernizer.md)** - Refactor legacy codebases and implement gradual modernization\n\n### ☁️ [Infrastructure](agents\u002Finfrastructure\u002F)\n\n- **[cloud-architect](agents\u002Finfrastructure\u002Fcloud-architect.md)** - Design AWS\u002FAzure\u002FGCP infrastructure and optimize cloud costs\n- **[deployment-engineer](agents\u002Finfrastructure\u002Fdeployment-engineer.md)** - Configure CI\u002FCD pipelines, Docker containers, and cloud deployments\n- **[devops-incident-responder](agents\u002Finfrastructure\u002Fdevops-incident-responder.md)** - Debug production issues, analyze logs, and fix deployment failures\n- **[incident-responder](agents\u002Finfrastructure\u002Fincident-responder.md)** - Handles production incidents with urgency and precision\n- **[performance-engineer](agents\u002Finfrastructure\u002Fperformance-engineer.md)** - Profile applications, optimize bottlenecks, and implement caching strategies\n\n### 🔍 [Quality & Testing](agents\u002Fquality-testing\u002F)\n\n- **[code-reviewer](agents\u002Fquality-testing\u002Fcode-reviewer.md)** - Expert code review for quality, security, and maintainability\n- **[architect-reviewer](agents\u002Fquality-testing\u002Farchitect-review.md)** - Reviews code changes for architectural consistency and design patterns\n- **[qa-expert](agents\u002Fquality-testing\u002Fqa-expert.md)** - Comprehensive QA processes and testing strategies for quality assurance\n- **[test-automator](agents\u002Fquality-testing\u002Ftest-automator.md)** - Create comprehensive test suites with unit, integration, and e2e tests\n- **[debugger](agents\u002Fquality-testing\u002Fdebugger.md)** - Debugging specialist for errors, test failures, and unexpected behavior\n\n### 📊 [Data & AI](agents\u002Fdata-ai\u002F)\n\n**Data Engineering & Analytics**\n\n- **[data-engineer](agents\u002Fdata-ai\u002Fdata-engineer.md)** - Build ETL pipelines, data warehouses, and streaming architectures\n- **[data-scientist](agents\u002Fdata-ai\u002Fdata-scientist.md)** - Data analysis expert for SQL queries, BigQuery operations, and data insights\n- **[database-optimizer](agents\u002Fdata-ai\u002Fdatabase-optimizer.md)** - Optimize SQL queries, design efficient indexes, and handle database migrations\n- **[postgres-pro](agents\u002Fdata-ai\u002Fpostgres-pro.md)** - PostgreSQL database expert for advanced queries and optimizations\n- **[graphql-architect](agents\u002Fdata-ai\u002Fgraphql-architect.md)** - Design GraphQL schemas, resolvers, and federation patterns\n\n**AI & Machine Learning**\n\n- **[ai-engineer](agents\u002Fdata-ai\u002Fai-engineer.md)** - Build LLM applications, RAG systems, and prompt pipelines\n- **[ml-engineer](agents\u002Fdata-ai\u002Fml-engineer.md)** - Implement ML pipelines, model serving, and feature engineering\n- **[prompt-engineer](agents\u002Fdata-ai\u002Fprompt-engineer.md)** - Optimizes prompts for LLMs and AI systems\n\n### 🛡️ [Security](agents\u002Fsecurity\u002F)\n\n- **[security-auditor](agents\u002Fsecurity\u002Fsecurity-auditor.md)** - Review code for vulnerabilities and ensure OWASP compliance\n\n### 🎯 [Specialization](agents\u002Fspecialization\u002F)\n\n- **[api-documenter](agents\u002Fspecialization\u002Fapi-documenter.md)** - Create OpenAPI\u002FSwagger specs and write developer documentation\n- **[documentation-expert](agents\u002Fspecialization\u002Fdocumentation-expert.md)** - Professional technical writing and comprehensive documentation systems\n\n### 💼 [Business](agents\u002Fbusiness\u002F)\n\n- **[product-manager](agents\u002Fbusiness\u002Fproduct-manager.md)** - Strategic product management with roadmap planning and stakeholder alignment\n\n### 🎭 Meta-Orchestration\n\n- **[agent-organizer](agents\u002Fagent-organizer.md)** - Master orchestrator for complex, multi-agent tasks. Analyzes project requirements, assembles optimal agent teams, and manages collaborative workflows for comprehensive project execution.\n\n**Key Capabilities:**\n\n- **Intelligent Project Analysis**: Technology stack detection, architecture pattern recognition, and requirement extraction\n- **Strategic Team Assembly**: Selects optimal 1-3 agent teams based on project needs and complexity\n- **Workflow Orchestration**: Manages multi-phase collaboration with quality gates and validation checkpoints\n- **Efficiency Optimization**: Focused teams for common tasks (bug fixes, features, documentation) with comprehensive orchestration for complex projects\n\n**When to Use**: Complex multi-step projects, cross-domain tasks, architecture decisions, comprehensive analysis, or any scenario requiring coordinated expertise from multiple specialized agents.\n\n## 📦 Installation\n\n### Quick Setup\n\n### Manual Installation (Recommend)\n\nAlternatively, you can manually copy individual agent files:\n\n```bash\n# Prevent replacing documents from other providers\nmkdir ~\u002F.claude\u002Fagents\u002Flst97\n# Copy specific agents to your Claude agents directory\ncp \u002Fpath\u002Fto\u002Fagents\u002F*.md ~\u002F.claude\u002Fagents\u002Flst97\n```\n\n### Verification\n\nTo verify agents are loaded correctly:\n\n```bash\n# List all available agents\nls ~\u002F.claude\u002Fagents\u002Flst97\u002F*.md\n\n# Check Claude Code recognizes the agents (run in Claude Code)\n# \"List all available subagents\"\n```\n\n### Quick Installation\n\nThese subagents are automatically available when placed in the `~\u002F.claude\u002Fagents\u002F` directory. Claude Code will automatically detect and load them on startup. This will enable the CLAUDE.md to be available in global scope, may also conflict with other repository.\n\n```bash\n# Clone the repository to your Claude agents directory\n# Documents are base on the scaffold from https:\u002F\u002Fgithub.com\u002Fwshobson\u002Fagents.git\ncd ~\u002F.claude\ngit clone https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents.git\n\n# Or if the directory already exists, pull the latest updates\ncd ~\u002F.claude\ngit pull origin main\n```\n\n### 🔧 MCP Server Configuration (Required for Full Performance)\n\nTo enable optimal performance with specialized MCP (Model Context Protocol) servers that enhance agent capabilities, add the following configuration to your **global** Claude settings file (`~\u002F.claude.json`):\n\n```json\n\"mcpServers\": {\n  \"sequential-thinking\": {\n    \"type\": \"stdio\",\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"@modelcontextprotocol\u002Fserver-sequential-thinking\"\n    ],\n    \"env\": {}\n  },\n  \"context7\": {\n    \"type\": \"stdio\",\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"@upstash\u002Fcontext7-mcp\"\n    ],\n    \"env\": {}\n  },\n  \"magic\": {\n    \"type\": \"stdio\",\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"@21st-dev\u002Fmagic@latest\",\n      \"API_KEY=\\\"api-key\\\"\" \u002F\u002F API key is required\n    ],\n    \"env\": {}\n  },\n  \"playwright\": {\n    \"type\": \"stdio\",\n    \"command\": \"npx\",\n    \"args\": [\n      \"@playwright\u002Fmcp@latest\"\n    ],\n    \"env\": {}\n  },\n  \"filesystem\": {\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"@modelcontextprotocol\u002Fserver-filesystem\",\n      \"\u002Fyour\u002Fallowed\u002Fpath\" \u002F\u002F please add your path here\n    ]\n  },\n  \"puppeteer\": {\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"puppeteer-mcp-server\"\n    ],\n    \"env\": {}\n  }\n}\n```\n\n**MCP Server Benefits:**\n\n- **sequential-thinking**: Enhanced multi-step reasoning and complex analysis\n- **context7**: Access to up-to-date documentation and framework patterns\n- **magic**: Advanced UI component generation and design system integration\n- **playwright**: Cross-browser testing and E2E automation capabilities\n\n**Note**: These MCP servers significantly enhance agent capabilities but are not strictly required for basic functionality.\n\n### 🎭 Advanced: Agent-Organizer Auto-Dispatch Setup\n\nFor complex projects requiring multi-agent coordination, you can enable the dispatch protocol in your **project root directory** (not globally):\n\n```bash\n# Copy CLAUDE.md to your PROJECT root directory (recommended)\ncp \u002Fpath\u002Fto\u002Fagents\u002FCLAUDE.md \u002Fpath\u002Fto\u002Fyour\u002Fproject\u002FCLAUDE.md\n```\n\n**⚠️ Project-Scope Recommendation:**\n\n- **✅ Project-Specific**: Place CLAUDE.md in individual project roots for targeted orchestration\n- **❌ Global Scope**: Avoid placing in `~\u002F.claude\u002FCLAUDE.md` to prevent over-orchestration of simple tasks\n- **🎯 Selective Usage**: Enable only for projects requiring comprehensive multi-agent workflows\n\n**Trade-offs to Consider:**\n\n- **Quality vs Speed**: Multi-agent workflows provide expert results but take longer\n- **Token Efficiency**: 2-5x token usage for comprehensive analysis and implementation\n- **Complexity Matching**: Best for complex projects, may over-engineer simple tasks\n\n## 🔧 Usage\n\n### Automatic Invocation (Recommended)\n\nClaude Code intelligently analyzes your request and automatically delegates to the most appropriate subagent(s) based on:\n\n- **Context Analysis**: Keywords, file types, and project structure\n- **Task Classification**: Development, debugging, optimization, etc.\n- **Domain Expertise**: Matching requirements to specialist knowledge\n- **Workflow Patterns**: Common multi-agent coordination scenarios\n\n**Example**: `\"Implement user authentication with secure password handling\"` → Automatically uses: `backend-architect` → `security-auditor` → `test-automator`\n\n### Explicit Invocation\n\nFor specific expertise or when you want control over agent selection:\n\n```bash\n# Direct agent requests\n\"Use the code-reviewer to check my recent changes\"\n\"Have the security-auditor scan for vulnerabilities\"\n\"Get the performance-engineer to optimize this bottleneck\"\n\n# Multi-agent requests\n\"Have backend-architect design the API, then security-auditor review it\"\n\"Use data-scientist to analyze this dataset, then ai-engineer to build recommendations\"\n```\n\n### Hybrid Approach\n\nCombine automatic and explicit invocation:\n\n```bash\n# Start explicit, let Claude coordinate the rest\n\"Use backend-architect to design a REST API for user management, then handle the implementation automatically\"\n\n# Explicit validation after automatic work\n\"Implement this feature automatically, then have security-auditor review the result\"\n```\n\n## 💡 Usage Examples\n\n### Direct Agent Invocation\n\nWhen not using agent-organizer, specify the exact agent needed for your task:\n\n```bash\n# Development Tasks\n\"Use backend-architect to design a REST API for user management\"\n\"Have frontend-developer create a responsive login form component\"\n\"Get python-pro to implement async data processing with proper error handling\"\n\"Have react-pro optimize this component for performance and add proper TypeScript types\"\n\"Use typescript-pro to refactor this module with advanced type safety\"\n\n# Code Quality & Review\n\"Use code-reviewer to analyze this pull request for best practices\"\n\"Have architect-reviewer check if this change maintains architectural consistency\"\n\"Get debugger to investigate why this test is failing intermittently\"\n\n# Security & Performance\n\"Have security-auditor scan this authentication module for vulnerabilities\"\n\"Use performance-engineer to identify bottlenecks in this API endpoint\"\n\"Get database-optimizer to improve these slow queries\"\n\n# Testing & QA\n\"Use test-automator to create comprehensive tests for this user service\"\n\"Have qa-expert design a testing strategy for this new feature\"\n\n# Infrastructure & Deployment\n\"Get devops-incident-responder to investigate this production deployment failure\"\n\"Use cloud-architect to design scalable infrastructure for this microservice\"\n\"Have deployment-engineer set up CI\u002FCD pipeline for this repository\"\n\n# Data & AI\n\"Use data-scientist to analyze user behavior patterns in this dataset\"\n\"Have ai-engineer implement a RAG system for document search\"\n\"Get ml-engineer to deploy this trained model to production\"\n\n# Documentation & Specialization\n\"Use documentation-expert to create comprehensive API documentation\"\n\"Have api-documenter generate OpenAPI specs for these endpoints\"\n\n# Multi-Agent Coordination Examples\n\"Use backend-architect to design the API, then have security-auditor review it\"\n\"Get frontend-developer to build the component, then use test-automator for coverage\"\n\"Have database-optimizer improve queries, then performance-engineer validate results\"\n```\n\n### Agent Communication Protocol Examples\n\nEach agent uses a standardized communication protocol with agent-specific context requests. Here are examples:\n\n#### Frontend Development\n\n```json\n{\n  \"requesting_agent\": \"frontend-developer\",\n  \"request_type\": \"get_task_briefing\",\n  \"payload\": {\n    \"query\": \"Initial briefing required for UI component development. Provide overview of existing React project structure, design system, component library, and relevant frontend files.\"\n  }\n}\n```\n\n## 📋 Subagent Format\n\nEach subagent follows a standardized structure for consistent behavior and optimal integration:\n\n### File Structure\n\n```markdown\n---\nname: subagent-name\ndescription: When this subagent should be invoked\ntools: tool1, tool2  # Optional - defaults to all tools\n---\n\n# Subagent Name\n\n**Role**: Detailed role description and primary responsibilities\n\n**Expertise**: Specific technologies, frameworks, and domain knowledge\n\n**Key Capabilities**:\n- Capability 1: Description\n- Capability 2: Description\n- Capability 3: Description\n\nSystem prompt defining the subagent's specialized behavior, decision-making patterns, and interaction style with other agents.\n```\n\n### Required Components\n\n- **Name**: Kebab-case filename matching the agent name\n- **Description**: Clear trigger conditions for automatic invocation\n- **Role Definition**: Specific responsibilities and boundaries\n- **Expertise Areas**: Technologies, patterns, and domain knowledge\n- **System Prompt**: Detailed instructions for specialized behavior\n\n### Optional Components\n\n- **Tools**: Specific Claude Code tools (defaults to all available tools)\n- **Dependencies**: Other agents this one commonly works with\n- **Patterns**: Common workflow patterns and coordination scenarios\n\n## 🔄 Agent Orchestration Patterns\n\nClaude Code automatically coordinates agents using these patterns:\n\n- **Sequential**: `architect → implement → test → review` for dependent tasks\n- **Parallel**: `performance-engineer + database-optimizer` for independent analysis  \n- **Validation**: `primary-agent → security-auditor` for critical components\n- **Iterative**: `review → refine → validate` for optimization tasks\n\n## 🎯 When to Use Which Agent\n\n### 🏗️ Planning & Architecture\n\n| Agent | Best For | Example Use Cases |\n|-------|----------|-------------------|\n| **[backend-architect](agents\u002Fdevelopment\u002Fbackend-architect.md)** | API design, system architecture | RESTful APIs, microservices, database schemas |\n| **[frontend-developer](agents\u002Fdevelopment\u002Ffrontend-developer.md)** | UI\u002FUX planning, component design | React components, responsive layouts, state management |\n| **[cloud-architect](agents\u002Finfrastructure\u002Fcloud-architect.md)** | Infrastructure design, scalability | AWS\u002FAzure\u002FGCP architecture, cost optimization |\n| **[graphql-architect](agents\u002Fdata-ai\u002Fgraphql-architect.md)** | GraphQL system design | Schema design, resolvers, federation |\n\n### 💻 Implementation & Development  \n\n| Agent | Best For | Example Use Cases |\n|-------|----------|-------------------|\n| **[python-pro](agents\u002Fdevelopment\u002Fpython-pro.md)** | Python development | Django\u002FFastAPI apps, data processing, async programming |\n| **[golang-pro](agents\u002Fdevelopment\u002Fgolang-pro.md)** | Go development | Microservices, concurrent systems, CLI tools |\n| **[typescript-pro](agents\u002Fdevelopment\u002Ftypescript-pro.md)** | TypeScript development | Type-safe applications, advanced TS features |\n| **[react-pro](agents\u002Fdevelopment\u002Freact-pro.md)** | React expertise | Hooks, performance optimization, advanced patterns |\n| **[nextjs-pro](agents\u002Fdevelopment\u002Fnextjs-pro.md)** | Next.js applications | SSR\u002FSSG, full-stack React, routing |\n\n### ☁️ Operations & Maintenance\n\n| Agent | Best For | Example Use Cases |\n|-------|----------|-------------------|\n| **[devops-incident-responder](agents\u002Finfrastructure\u002Fdevops-incident-responder.md)** | Production issues, deployments | Log analysis, deployment failures, system debugging |\n| **[incident-responder](agents\u002Finfrastructure\u002Fincident-responder.md)** | Critical outages | Immediate response, crisis management, escalation |\n| **[deployment-engineer](agents\u002Finfrastructure\u002Fdeployment-engineer.md)** | CI\u002FCD, containerization | Docker, Kubernetes, pipeline configuration |\n| **[database-optimizer](agents\u002Fdata-ai\u002Fdatabase-optimizer.md)** | Database performance | Query optimization, indexing, migration strategies |\n\n### 📊 Analysis & Optimization\n\n| Agent | Best For | Example Use Cases |\n|-------|----------|-------------------|\n| **[performance-engineer](agents\u002Finfrastructure\u002Fperformance-engineer.md)** | Application performance | Bottleneck analysis, caching strategies, optimization |\n| **[security-auditor](agents\u002Fsecurity\u002Fsecurity-auditor.md)** | Security assessment | Vulnerability scanning, OWASP compliance, threat modeling |\n| **[data-scientist](agents\u002Fdata-ai\u002Fdata-scientist.md)** | Data analysis | SQL queries, BigQuery, insights and reporting |\n| **[code-reviewer](agents\u002Fquality-testing\u002Fcode-reviewer.md)** | Code quality | Best practices, maintainability, architectural review |\n\n### 🧪 Quality Assurance\n\n| Agent | Best For | Example Use Cases |\n|-------|----------|-------------------|\n| **[test-automator](agents\u002Fquality-testing\u002Ftest-automator.md)** | Testing strategy | Unit tests, integration tests, E2E test suites |\n| **[debugger](agents\u002Fquality-testing\u002Fdebugger.md)** | Bug investigation | Error analysis, test failures, troubleshooting |\n| **[architect-reviewer](agents\u002Fquality-testing\u002Farchitect-review.md)** | Design validation | Architectural consistency, pattern compliance |\n\n## 📚 Best Practices\n\n- **Trust Auto-Delegation**: Claude Code excels at context analysis and optimal agent selection\n- **Provide Rich Context**: Include tech stack, constraints, and project background\n- **Use Explicit Control**: Override automatic selection when you need specific expertise\n- **Establish Quality Gates**: Build review and validation into standard workflows\n- **Match Task Complexity**: Don't over-engineer simple tasks or under-resource complex ones\n\n## 🤝 Contributing\n\n### Adding New Agents\n\nTo contribute a new subagent to the collection:\n\n1. **Follow Naming Convention**\n   - Use lowercase, hyphen-separated names (e.g., `backend-architect.md`)\n   - Name should clearly indicate the agent's domain and role\n\n2. **Use Standard Format**\n   - Include proper frontmatter with `name`, `description`, and optional `tools`\n   - Follow the structured format outlined in the [Subagent Format](#-subagent-format) section\n\n3. **Write Clear Descriptions**\n   - Description should clearly indicate when the agent should be automatically invoked\n   - Include specific keywords and contexts that trigger the agent\n\n4. **Define Specialized Behavior**\n   - Include detailed system prompt with role, expertise, and capabilities\n   - Define interaction patterns with other agents\n   - Specify decision-making frameworks and priorities\n\n5. **Test Integration**\n   - Verify the agent can be automatically invoked based on description\n   - Test explicit invocation with clear requests\n   - Ensure compatibility with existing agent coordination patterns\n\n### Quality Standards\n\n- **Domain Expertise**: Agents should demonstrate deep knowledge in their specialization\n- **Clear Boundaries**: Define what the agent does and doesn't handle\n- **Integration Ready**: Design for seamless coordination with other agents\n- **Consistent Voice**: Maintain professional, helpful, and expert tone\n\n### Submission Process\n\n1. Create the agent file following all standards\n2. Test the agent with various invocation patterns\n3. Submit a pull request with example use cases\n4. Include documentation of the agent's unique value and integration patterns\n\n## 🛠️ Troubleshooting\n\n**Common Issues:**\n\n- **Agent not selected**: Use domain-specific keywords or explicit invocation\n- **Unexpected selection**: Provide more context about tech stack and requirements\n- **Generic responses**: Request specific depth and include detailed constraints\n- **Conflicting advice**: Request reconciliation between different specialists\n\n**Resources:**\n\n- [Claude Code Documentation](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) - Official guide\n- [Subagents Documentation](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code\u002Fsub-agents) - Agent system reference\n\n## 📊 Quick Reference\n\n### Most Commonly Used Agents\n\n1. **[code-reviewer](agents\u002Fquality-testing\u002Fcode-reviewer.md)** - Quality assurance and best practices\n2. **[backend-architect](dagents\u002Fevelopment\u002Fbackend-architect.md)** - API and system design\n3. **[frontend-developer](agents\u002Fdevelopment\u002Ffrontend-developer.md)** - UI\u002FUX implementation\n4. **[security-auditor](agents\u002Fsecurity\u002Fsecurity-auditor.md)** - Security validation and compliance\n5. **[performance-engineer](agents\u002Finfrastructure\u002Fperformance-engineer.md)** - Optimization and bottleneck analysis\n\n### Essential Coordination Patterns\n\n- **Development**: `architect → implement → test → review`\n- **Debugging**: `debugger → specialist → validator`\n- **Optimization**: `performance-engineer + database-optimizer → validation`\n- **Security**: `primary-agent → security-auditor → approval`\n\n### Key Success Factors\n\n- ✅ Trust automatic delegation for optimal results\n- ✅ Provide rich context and specific requirements\n- ✅ Use explicit invocation strategically\n- ✅ Establish quality gates and validation patterns\n- ✅ Learn from agent coordination patterns\n\n## 🎬 Examples\n\nThese examples demonstrate real-world multi-agent coordination scenarios with detailed resource metrics to help you understand the token usage, execution time, and expected deliverables for different project complexities:\n\n- **Example 1**: Simple feature implementation (~300K tokens, ~17 minutes) - Shows efficient 4-agent coordination for focused component development\n- **Example 2**: Complex system implementation (~850K tokens, ~45 minutes) - Demonstrates enterprise-scale 7-agent orchestration with error recovery\n\nBoth examples include actual token counts, execution times, and deliverable quality to help you plan and budget for multi-agent workflows in your projects.\n\n### Example 1: ExportStep Component Implementation\n\n**User Request**: `\u002Fsc:implement` use agent-organizer to design and implement the ExportStep.tsx function, improve the UI\u002FUX as well\n\n#### Agent Orchestration Flow\n\n![Agent Orchestration](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_52f11606211e.png)\n\n**Step 1: agent-organizer Analysis** (56.7K tokens, 1m 20s)\n\n- Analyzed existing project structure and Zustand stores\n- Created comprehensive 7-task implementation plan\n- Assembled 3-agent specialist team for coordinated execution\n\n![Backend Implementation](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_f75b3e29f83c.png)\n\n**Step 2: backend-architect Implementation** (99.1K tokens, 7m 31s)\n\n- Designed complete export store architecture with state management\n- Implemented format conversion utilities for SRT, VTT, ASS, JSON formats\n- Integrated Electron IPC for seamless file operations\n\n![Frontend Enhancement](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_abd528092263.png)\n\n**Step 3: frontend-developer Enhancement** (84.3K tokens, 5m 29s)\n\n- Created fully interactive ExportStep component with real event handlers\n- Implemented real-time preview generation based on subtitle data\n- Added accessibility compliance (WCAG 2.1 AA) and responsive design\n\n![Testing Strategy](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_fc90f9d581c7.png)\n\n**Step 4: test-automator Quality Assurance** (61.4K tokens, 2m 46s)\n\n- Developed comprehensive test coverage for format converters\n- Set up Jest and React Testing Library framework\n- Created accessibility and interaction testing strategies\n\n#### Implementation Results\n\n![Final Output](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_746297a8803a.png)\n\n**Complete Feature Delivery**:\n\n- 🏗️ **Backend**: Export store with state management, format conversion utilities, Electron IPC integration\n- 🎨 **Frontend**: Interactive UI with real-time preview, accessibility compliance, keyboard navigation\n- ✅ **Testing**: Comprehensive test coverage with framework setup and validation\n\n![Live Demo](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_877b30cc06c5.gif)\n\n#### Project Metrics\n\n**Resource Usage**:\n\n- **Total Tokens**: ~301K tokens (agent-organizer: 56K, backend-architect: 99K, frontend-developer: 84K, test-automator: 61K)\n- **Total Time**: ~30 minutes execution time\n- **Team Size**: 4 agents (1 orchestrator + 3 specialists)\n- **Files Created\u002FModified**: 4 major files (stores, components, utilities, tests)\n\n**Efficiency Highlights**:\n\n- **Sequential Coordination**: Each agent built upon previous work seamlessly\n- **Quality Integration**: Production-ready export system with comprehensive functionality\n- **Zero Breaking Changes**: Enhanced existing architecture without disruption\n\n### Example 2: Complex Workspace Management System\n\n**User Request**: `\u002Fsc:design` implement complex workspace management with user config persistence, multiple workspaces, workspace groups, Discord-like UI with drag-and-drop functionality\n\n#### Phase 1: Comprehensive Design & Multi-Agent Assessment\n\n![Agent Organizer Design Phase](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_16ba993d2001.png)\n\n**5-Agent Team Assembly**: backend-architect, frontend-developer, electron-pro, ux-designer, test-automator\n\n**Design Deliverables**:\n\n- Complete TypeScript interfaces for Workspace, WorkspaceGroup, and configurations\n- IndexedDB storage strategy with migration from localStorage  \n- Discord-inspired UI specifications with drag-and-drop functionality\n- Auto-save mechanisms with conflict resolution and backup strategy\n- 5-phase implementation plan with quality gates\n\n![Phase 1 Working](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_c683a9037368.png)\n\n**Phase 1 Assessment Results**:\n\n![Phase 1 Complete](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_40e572e12c7c.png)\n![Phase 1 Summary](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_d472a2472888.png)\n\n**Comprehensive Team Assessment** (5 agents, ~400K tokens total):\n\n- 🏗️ **Backend Architecture**: IndexedDB schema, \u003C200ms startup, migration framework, auto-save strategy\n- 🎨 **Frontend Components**: Discord-inspired design, Material-UI integration, progressive enhancement\n- ⚡ **Electron Integration**: IPC architecture, security model, performance optimization\n- 🎭 **UX Design**: A+ UX Score (92\u002F100), zero disruption, user journey validation  \n- ✅ **Testing Strategy**: 99.5% migration success, 4-layer testing pyramid, quality gates\n\n#### Complete Implementation Results\n\n![All Phases Complete](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_13c183c3e288.png)\n\n**Full 5-Phase Implementation**:\n\n- **Phase 1**: Assessment & Current State Analysis ✅\n- **Phase 2**: Architecture Finalization & Infrastructure ✅  \n- **Phase 3**: Core Implementation ✅\n- **Phase 4**: Integration & Migration ✅\n- **Phase 5**: Quality Assurance & Finalization ✅\n\n**Final Deliverables**:\n\n- Complete workspace management system with IndexedDB persistence\n- Discord-inspired UI with drag-and-drop workspace organization\n- Multi-workspace support with workspace groups\n- Seamless migration from existing localStorage system\n- Comprehensive test coverage and error recovery mechanisms\n\n#### Resource Metrics & Performance\n\n**Total Project Metrics**:\n\n- **Tokens Used**: ~900K tokens across all phases and error resolution\n- **Time Spent**: ~120 minutes total execution time\n- **Agents Involved**: 7 specialized agents (5 primary + 2 error resolution)\n- **Lines of Code**: ~2,400 lines across 15+ files\n- **Test Coverage**: 99.5% with comprehensive edge case handling (Should be hallucination)\n\n#### Build Error Resolution with Nested Agent Coordination\n\n![Build Error Detection](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_aaf81763fef8.png)\n\n**Second User Prompt**: `@agent-code-reviewer-pro` the application have build error please find all the build errors and ask the related sub agent to fix it. `@agent-agent-organizer`\n\n![Nested Sub-Agent Coordination](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_f3d4ad709768.png)\n\n**Error Resolution Flow**:\n\n1. **code-reviewer-pro** (68.5K tokens, 5m 26s): Identified critical TypeScript syntax errors\n2. **agent-organizer** coordination: Systematic build error fixes with **typescript-pro**\n3. **Nested delegation**: Specialized agents called within agent workflows for targeted fixes\n\n**Error Resolution Efficiency**:\n\n- **Detection**: ~5m with code-reviewer-pro\n- **Coordination**: Instant agent-organizer response\n- **Fix Implementation**: ~30m minutes with nested typescript-pro agent\n- **Build Success**: Zero remaining errors after systematic fixes\n- **Challenging Runtime ERROR** Runtime error occur and it require manuel debugging and instruction\n\n### Key Multi-Agent Benefits\n\n- **🧠 Intelligent Orchestration**: agent-organizer coordinated 5+ agents across complex 5-phase implementation\n- **🔧 Nested Agent Support**: Error resolution through coordinated sub-agent delegation within workflows  \n- **📊 Enterprise-Scale Quality**: 850K tokens of comprehensive analysis, design, and implementation\n- **⚡ Rapid Error Recovery**: Build errors resolved in \u003C8 minutes through specialized agent coordination\n- **🎯 Domain Expertise**: Each agent contributed specialized knowledge (storage architecture, UX design, TypeScript fixes)\n\n---\n\n*Happy coding with your AI specialist team! 🚀*\n","# Claude Code 子代理集合\n\n面向 [Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) 的全面子代理集合，包含 33 个专业子代理，旨在通过领域专业知识和智能自动化提升开发工作流效率。\n\n## 🚀 概述\n\n本仓库收录了一系列精心挑选的专业子代理，可扩展 Claude Code 在整个软件开发生命周期中的能力。每个子代理专注于特定领域，可根据上下文分析自动调用，或在需要专业技能时被显式调用。\n\n### 核心特性\n\n- **🤖 智能自动委派**：Claude Code 会根据任务上下文自动选择最优代理。\n- **🔧 领域专长**：每个代理专注于特定技术、模式和最佳实践。\n- **🔄 多代理编排**：代理间无缝协作，支持复杂工作流。\n- **📊 质量保证**：所有领域均内置评审与验证机制。\n- **⚡ 性能优化**：代理设计注重高效完成任务及资源利用率。\n\n## 可用子代理\n\n现将代理按逻辑类别组织，便于导航：\n\n### 🏗️ [开发](agents\u002Fdevelopment\u002F)\n\n**前端与 UI 专家**\n\n- **[frontend-developer](agents\u002Fdevelopment\u002Ffrontend-developer.md)** - 构建 React 组件、实现响应式布局并管理客户端状态。\n- **[ui-designer](agents\u002Fdevelopment\u002Fui-designer.md)** - 致力于用户友好界面的创意设计。\n- **[ux-designer](agents\u002Fdevelopment\u002Fux-designer.md)** - 专注于用户体验设计与交互优化。\n- **[react-pro](agents\u002Fdevelopment\u002Freact-pro.md)** - 精通 React 开发，包括 Hooks、性能优化及最佳实践。\n- **[nextjs-pro](agents\u002Fdevelopment\u002Fnextjs-pro.md)** - Next.js 专家，擅长服务端渲染 (SSR)、静态生成 (SSG) 以及全栈 React 应用开发。\n\n**后端与架构**\n\n- **[backend-architect](agents\u002Fdevelopment\u002Fbackend-architect.md)** - 设计 RESTful API、微服务边界及数据库模式。\n- **[full-stack-developer](agents\u002Fdevelopment\u002Ffull-stack-developer.md)** - 实现从 UI 到数据库的端到端 Web 应用开发，并确保无缝集成。\n\n**语言专家**\n\n- **[python-pro](agents\u002Fdevelopment\u002Fpython-pro.md)** - 编写符合 Python 语言习惯的代码，兼具高级特性和优化。\n- **[golang-pro](agents\u002Fdevelopment\u002Fgolang-pro.md)** - 编写规范的 Go 代码，熟练运用协程、通道和接口。\n- **[typescript-pro](agents\u002Fdevelopment\u002Ftypescript-pro.md)** - 进行高级 TypeScript 开发，保障类型安全并采用现代模式。\n\n**平台与移动端**\n\n- **[mobile-developer](agents\u002Fdevelopment\u002Fmobile-developer.md)** - 使用 React Native 或 Flutter 开发具备原生集成的应用。\n- **[electron-pro](agents\u002Fdevelopment\u002Felectorn-pro.md)** - 专注 Electron 桌面应用开发及跨平台解决方案。\n\n**开发者体验**\n\n- **[dx-optimizer](agents\u002Fdevelopment\u002Fdx-optimizer.md)** - 专注于改善开发工具、环境搭建及工作流程的专家。\n- **[legacy-modernizer](agents\u002Fdevelopment\u002Flegacy-modernizer.md)** - 重构遗留代码库，逐步推进现代化改造。\n\n### ☁️ [基础设施](agents\u002Finfrastructure\u002F)\n\n- **[cloud-architect](agents\u002Finfrastructure\u002Fcloud-architect.md)** - 设计 AWS\u002FAzure\u002FGCP 基础设施，并优化云成本。\n- **[deployment-engineer](agents\u002Finfrastructure\u002Fdeployment-engineer.md)** - 配置 CI\u002FCD 流水线、Docker 容器及云端部署。\n- **[devops-incident-responder](agents\u002Finfrastructure\u002Fdevops-incident-responder.md)** - 排查生产环境问题、分析日志并修复部署故障。\n- **[incident-responder](agents\u002Finfrastructure\u002Fincident-responder.md)** - 以紧急且精准的方式处理生产事故。\n- **[performance-engineer](agents\u002Finfrastructure\u002Fperformance-engineer.md)** - 对应用进行性能剖析、优化瓶颈并实施缓存策略。\n\n### 🔍 [质量与测试](agents\u002Fquality-testing\u002F)\n\n- **[code-reviewer](agents\u002Fquality-testing\u002Fcode-reviewer.md)** - 专注于代码质量、安全性及可维护性的专家级评审。\n- **[architect-reviewer](agents\u002Fquality-testing\u002Farchitect-review.md)** - 审核代码变更是否符合架构一致性与设计模式。\n- **[qa-expert](agents\u002Fquality-testing\u002Fqa-expert.md)** - 提供全面的 QA 流程与测试策略，确保质量。\n- **[test-automator](agents\u002Fquality-testing\u002Ftest-automator.md)** - 创建涵盖单元测试、集成测试及端到端测试的完整测试套件。\n- **[debugger](agents\u002Fquality-testing\u002Fdebugger.md)** - 专门解决错误、测试失败及异常行为的调试专家。\n\n### 📊 [数据与 AI](agents\u002Fdata-ai\u002F)\n\n**数据工程与分析**\n\n- **[data-engineer](agents\u002Fdata-ai\u002Fdata-engineer.md)** - 构建 ETL 流水线、数据仓库及流式架构。\n- **[data-scientist](agents\u002Fdata-ai\u002Fdata-scientist.md)** - 精通 SQL 查询、BigQuery 操作及数据洞察的数据分析专家。\n- **[database-optimizer](agents\u002Fdata-ai\u002Fdatabase-optimizer.md)** - 优化 SQL 查询、设计高效索引并处理数据库迁移。\n- **[postgres-pro](agents\u002Fdata-ai\u002Fpostgres-pro.md)** - PostgreSQL 数据库专家，擅长高级查询与优化。\n- **[graphql-architect](agents\u002Fdata-ai\u002Fgraphql-architect.md)** - 设计 GraphQL 模式、解析器及联邦模式。\n\n**AI 与机器学习**\n\n- **[ai-engineer](agents\u002Fdata-ai\u002Fai-engineer.md)** - 构建 LLM 应用、RAG 系统及提示词流水线。\n- **[ml-engineer](agents\u002Fdata-ai\u002Fml-engineer.md)** - 实现机器学习流水线、模型推理及特征工程。\n- **[prompt-engineer](agents\u002Fdata-ai\u002Fprompt-engineer.md)** - 优化 LLM 和 AI 系统的提示词。\n\n### 🛡️ [安全](agents\u002Fsecurity\u002F)\n\n- **[security-auditor](agents\u002Fsecurity\u002Fsecurity-auditor.md)** - 审查代码漏洞并确保符合 OWASP 标准。\n\n### 🎯 [专业化](agents\u002Fspecialization\u002F)\n\n- **[api-documenter](agents\u002Fspecialization\u002Fapi-documenter.md)** - 生成 OpenAPI\u002FSwagger 规范并撰写开发者文档。\n- **[documentation-expert](agents\u002Fspecialization\u002Fdocumentation-expert.md)** - 专业的技术写作与完善的文档体系。\n\n### 💼 [业务](agents\u002Fbusiness\u002F)\n\n- **[product-manager](agents\u002Fbusiness\u002Fproduct-manager.md)** - 具备战略眼光的产品经理，负责制定产品路线图并与利益相关方协调。\n\n### 🎭 元编排\n\n- **[agent-organizer](agents\u002Fagent-organizer.md)** - 复杂多智能体任务的主控编排器。分析项目需求，组建最优智能体团队，并管理协作工作流以实现全面的项目执行。\n\n**核心能力：**\n\n- **智能项目分析**：技术栈检测、架构模式识别与需求提取\n- **战略团队组建**：根据项目需求和复杂度选择1至3个最优智能体团队\n- **工作流编排**：管理包含质量门和验证检查点的多阶段协作流程\n- **效率优化**：针对常见任务（Bug修复、功能开发、文档编写）组建专注团队，同时为复杂项目提供全面编排支持\n\n**适用场景**：复杂的多步骤项目、跨领域任务、架构决策、全面分析，或任何需要多个专业智能体协同工作的场景。\n\n## 📦 安装\n\n### 快速安装\n\n### 手动安装（推荐）\n\n您也可以手动复制单个智能体文件：\n\n```bash\n# 防止覆盖其他提供商的文档\nmkdir ~\u002F.claude\u002Fagents\u002Flst97\n# 将特定智能体复制到您的Claude智能体目录\ncp \u002Fpath\u002Fto\u002Fagents\u002F*.md ~\u002F.claude\u002Fagents\u002Flst97\n```\n\n### 验证\n\n要验证智能体是否正确加载：\n\n```bash\n# 列出所有可用智能体\nls ~\u002F.claude\u002Fagents\u002Flst97\u002F*.md\n\n# 检查Claude Code是否识别这些智能体（在Claude Code中运行）\n# “列出所有可用子智能体”\n```\n\n### 快速安装\n\n将这些子智能体放置在`~\u002F.claude\u002Fagents\u002F`目录下即可自动使用。Claude Code会在启动时自动检测并加载它们。这会使CLAUDE.md在全局范围内可用，但也可能与其他仓库发生冲突。\n\n```bash\n# 将仓库克隆到您的Claude智能体目录\n# 文档基于https:\u002F\u002Fgithub.com\u002Fwshobson\u002Fagents.git的脚手架\ncd ~\u002F.claude\ngit clone https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents.git\n\n# 或者如果目录已存在，拉取最新更新\ncd ~\u002F.claude\ngit pull origin main\n```\n\n### 🔧 MCP服务器配置（实现最佳性能所需）\n\n为了在专用MCP（模型上下文协议）服务器上实现最佳性能，从而增强智能体的能力，您需要在Claude的**全局**设置文件（`~\u002F.claude.json`）中添加以下配置：\n\n```json\n\"mcpServers\": {\n  \"sequential-thinking\": {\n    \"type\": \"stdio\",\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"@modelcontextprotocol\u002Fserver-sequential-thinking\"\n    ],\n    \"env\": {}\n  },\n  \"context7\": {\n    \"type\": \"stdio\",\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"@upstash\u002Fcontext7-mcp\"\n    ],\n    \"env\": {}\n  },\n  \"magic\": {\n    \"type\": \"stdio\",\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"@21st-dev\u002Fmagic@latest\",\n      \"API_KEY=\\\"api-key\\\"\" \u002F\u002F 需要API密钥\n    ],\n    \"env\": {}\n  },\n  \"playwright\": {\n    \"type\": \"stdio\",\n    \"command\": \"npx\",\n    \"args\": [\n      \"@playwright\u002Fmcp@latest\"\n    ],\n    \"env\": {}\n  },\n  \"filesystem\": {\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"@modelcontextprotocol\u002Fserver-filesystem\",\n      \"\u002Fyour\u002Fallowed\u002Fpath\" \u002F\u002F 请在此处添加您的路径\n    ]\n  },\n  \"puppeteer\": {\n    \"command\": \"npx\",\n    \"args\": [\n      \"-y\",\n      \"puppeteer-mcp-server\"\n    ],\n    \"env\": {}\n  }\n}\n```\n\n**MCP服务器的优势：**\n\n- **sequential-thinking**：增强多步推理与复杂分析能力\n- **context7**：访问最新文档和框架模式\n- **magic**：高级UI组件生成与设计系统集成\n- **playwright**：跨浏览器测试与端到端自动化能力\n\n**注意**：这些MCP服务器能显著提升智能体能力，但并非基本功能所必需。\n\n### 🎭 高级：Agent-Organizer自动调度设置\n\n对于需要多智能体协调的复杂项目，您可以在**项目根目录**（而非全局）启用调度协议：\n\n```bash\n# 将CLAUDE.md复制到您的项目根目录（推荐）\ncp \u002Fpath\u002Fto\u002Fagents\u002FCLAUDE.md \u002Fpath\u002Fto\u002Fyour\u002Fproject\u002FCLAUDE.md\n```\n\n**⚠️ 项目范围建议：**\n\n- **✅ 项目专用**：将CLAUDE.md放置在各个项目根目录中，以实现针对性的编排\n- **❌ 全局范围**：避免将其放置在`~\u002F.claude\u002FCLAUDE.md`中，以免对简单任务过度编排\n- **🎯 选择性使用**：仅在需要全面多智能体工作流的项目中启用\n\n**需权衡的方面：**\n\n- **质量 vs 速度**：多智能体工作流可提供专家级结果，但耗时较长\n- **Token效率**：进行全面分析和实施时，Token消耗会增加2至5倍\n- **复杂度匹配**：最适合复杂项目，可能会使简单任务过度工程化\n\n## 🔧 使用方法\n\n### 自动调用（推荐）\n\nClaude Code会智能分析您的请求，并根据以下因素自动委派给最合适的子智能体：\n\n- **上下文分析**：关键词、文件类型及项目结构\n- **任务分类**：开发、调试、优化等\n- **领域专长**：将需求与专业知识匹配\n- **工作流模式**：常见的多智能体协调场景\n\n**示例**：“实现用户认证，并确保密码安全处理” → 自动使用：`backend-architect` → `security-auditor` → `test-automator`\n\n### 显式调用\n\n当需要特定专业知识或希望控制智能体选择时：\n\n```bash\n# 直接请求智能体\n“请code-reviewer检查我最近的更改”\n“让security-auditor扫描漏洞”\n“请performance-engineer优化这个瓶颈”\n\n# 多智能体请求\n“先让backend-architect设计API，再由security-auditor审核”\n“用data-scientist分析这个数据集，然后让ai-engineer构建推荐系统”\n```\n\n### 混合方式\n\n结合自动与显式调用：\n\n```bash\n# 显式开始，让Claude协调剩余部分\n“先用backend-architect设计用户管理的REST API，然后自动完成实现”\n\n# 在自动完成后进行显式验证\n“自动实现这个功能，然后让security-auditor审查结果”\n```\n\n## 💡 使用示例\n\n### 直接调用智能体\n\n在不使用agent-organizer时，明确指定完成任务所需的智能体：\n\n```bash\n# 开发任务\n“用backend-architect设计用户管理的REST API”\n“让frontend-developer创建响应式的登录表单组件”\n“请python-pro实现异步数据处理，并做好错误处理”\n“让react-pro优化该组件的性能，同时添加正确的TypeScript类型”\n“用typescript-pro重构此模块，提升类型安全性”\n\n# 代码质量与评审\n“使用代码评审员分析此拉取请求的最佳实践”\n“请架构师评审员检查此更改是否保持架构一致性”\n“让调试器调查此测试为何间歇性失败”\n\n# 安全与性能\n“请安全审计员扫描此认证模块是否存在漏洞”\n“使用性能工程师识别此 API 端点中的瓶颈”\n“让数据库优化师优化这些慢查询”\n\n# 测试与 QA\n“使用测试自动化工具为此用户服务创建全面的测试”\n“请 QA 专家为此新功能设计测试策略”\n\n# 基础设施与部署\n“请 DevOps 事件响应人员调查此次生产部署失败的原因”\n“使用云架构师为此微服务设计可扩展的基础设施”\n“请部署工程师为此仓库搭建 CI\u002FCD 流水线”\n\n# 数据与 AI\n“使用数据科学家分析此数据集中的用户行为模式”\n“请 AI 工程师实现用于文档搜索的 RAG 系统”\n“让机器学习工程师将此训练好的模型部署到生产环境”\n\n# 文档与专业化\n“使用文档专家创建全面的 API 文档”\n“请 API 文档生成器为这些端点生成 OpenAPI 规范”\n\n# 多智能体协调示例\n“使用后端架构师设计 API，然后请安全审计员进行审查”\n“让前端开发者构建组件，再使用测试自动化工具确保覆盖率”\n“请数据库优化师优化查询，再由性能工程师验证结果”\n```\n\n### 智能体通信协议示例\n\n每个智能体都使用带有特定于智能体上下文请求的标准化通信协议。以下是示例：\n\n#### 前端开发\n\n```json\n{\n  \"requesting_agent\": \"frontend-developer\",\n  \"request_type\": \"get_task_briefing\",\n  \"payload\": {\n    \"query\": \"需要 UI 组件开发的初始简报。请提供现有 React 项目结构、设计系统、组件库以及相关前端文件的概述。\"\n  }\n}\n```\n\n## 📋 子智能体格式\n\n每个子智能体都遵循标准化结构，以确保一致的行为和最佳集成：\n\n### 文件结构\n\n```markdown\n---\nname: subagent-name\ndescription: 何时应调用此子智能体\ntools: tool1, tool2  # 可选 - 默认为所有工具\n---\n\n# 子智能体名称\n\n**角色**: 详细的角色描述及主要职责\n\n**专业知识**: 具体的技术、框架和领域知识\n\n**关键能力**:\n- 能力 1: 描述\n- 能力 2: 描述\n- 能力 3: 描述\n\n定义子智能体专业行为、决策模式及其与其他智能体交互方式的系统提示。\n```\n\n### 必需组件\n\n- **名称**: 与智能体名称匹配的短横线分隔文件名\n- **描述**: 明确自动调用的触发条件\n- **角色定义**: 具体的责任范围和边界\n- **专业知识领域**: 技术、模式和领域知识\n- **系统提示**: 详细的专业行为指令\n\n### 可选组件\n\n- **工具**: 特定的 Claude Code 工具（默认为所有可用工具）\n- **依赖关系**: 此智能体常合作的其他智能体\n- **模式**: 常见的工作流程模式和协作场景\n\n## 🔄 智能体编排模式\n\nClaude Code 自动使用以下模式协调智能体：\n\n- **顺序模式**: `architect → implement → test → review` 用于依赖性任务\n- **并行模式**: `performance-engineer + database-optimizer` 用于独立分析\n- **验证模式**: `primary-agent → security-auditor` 用于关键组件\n- **迭代模式**: `review → refine → validate` 用于优化任务\n\n## 🎯 何时使用哪个智能体\n\n### 🏗️ 规划与架构\n\n| 智能体 | 最适合 | 示例用例 |\n|-------|----------|-------------------|\n| **[backend-architect](agents\u002Fdevelopment\u002Fbackend-architect.md)** | API 设计、系统架构 | RESTful API、微服务、数据库模式 |\n| **[frontend-developer](agents\u002Fdevelopment\u002Ffrontend-developer.md)** | UI\u002FUX 规划、组件设计 | React 组件、响应式布局、状态管理 |\n| **[cloud-architect](agents\u002Finfrastructure\u002Fcloud-architect.md)** | 基础设施设计、可扩展性 | AWS\u002FAzure\u002FGCP 架构、成本优化 |\n| **[graphql-architect](agents\u002Fdata-ai\u002Fgraphql-architect.md)** | GraphQL 系统设计 | 模式设计、解析器、联邦 |\n\n### 💻 实施与开发\n\n| 智能体 | 最适合 | 示例用例 |\n|-------|----------|-------------------|\n| **[python-pro](agents\u002Fdevelopment\u002Fpython-pro.md)** | Python 开发 | Django\u002FFastAPI 应用、数据处理、异步编程 |\n| **[golang-pro](agents\u002Fdevelopment\u002Fgolang-pro.md)** | Go 开发 | 微服务、并发系统、CLI 工具 |\n| **[typescript-pro](agents\u002Fdevelopment\u002Ftypescript-pro.md)** | TypeScript 开发 | 类型安全的应用、高级 TypeScript 特性 |\n| **[react-pro](agents\u002Fdevelopment\u002Freact-pro.md)** | React 专长 | Hooks、性能优化、高级模式 |\n| **[nextjs-pro](agents\u002Fdevelopment\u002Fnextjs-pro.md)** | Next.js 应用 | SSR\u002FSSG、全栈 React、路由 |\n\n### ☁️ 运营与维护\n\n| 智能体 | 最适合 | 示例用例 |\n|-------|----------|-------------------|\n| **[devops-incident-responder](agents\u002Finfrastructure\u002Fdevops-incident-responder.md)** | 生产问题、部署 | 日志分析、部署失败、系统调试 |\n| **[incident-responder](agents\u002Finfrastructure\u002Fincident-responder.md)** | 重大故障 | 紧急响应、危机管理、升级处理 |\n| **[deployment-engineer](agents\u002Finfrastructure\u002Fdeployment-engineer.md)** | CI\u002FCD、容器化 | Docker、Kubernetes、流水线配置 |\n| **[database-optimizer](agents\u002Fdata-ai\u002Fdatabase-optimizer.md)** | 数据库性能 | 查询优化、索引创建、迁移策略 |\n\n### 📊 分析与优化\n\n| 智能体 | 最适合 | 示例用例 |\n|-------|----------|-------------------|\n| **[performance-engineer](agents\u002Finfrastructure\u002Fperformance-engineer.md)** | 应用性能 | 瓶颈分析、缓存策略、优化 |\n| **[security-auditor](agents\u002Fsecurity\u002Fsecurity-auditor.md)** | 安全评估 | 漏洞扫描、OWASP 合规、威胁建模 |\n| **[data-scientist](agents\u002Fdata-ai\u002Fdata-scientist.md)** | 数据分析 | SQL 查询、BigQuery、洞察与报告 |\n| **[code-reviewer](agents\u002Fquality-testing\u002Fcode-reviewer.md)** | 代码质量 | 最佳实践、可维护性、架构审查 |\n\n### 🧪 质量保证\n\n| 代理 | 最适合 | 示例用例 |\n|-------|----------|-------------------|\n| **[test-automator](agents\u002Fquality-testing\u002Ftest-automator.md)** | 测试策略 | 单元测试、集成测试、端到端测试套件 |\n| **[debugger](agents\u002Fquality-testing\u002Fdebugger.md)** | 错误排查 | 错误分析、测试失败、故障排除 |\n| **[architect-reviewer](agents\u002Fquality-testing\u002Farchitect-review.md)** | 设计验证 | 架构一致性、模式合规性 |\n\n## 📚 最佳实践\n\n- **信任自动委派**：Claude Code 在上下文分析和最优代理选择方面表现出色\n- **提供丰富上下文**：包括技术栈、约束条件和项目背景\n- **使用显式控制**：当需要特定专业知识时，可覆盖自动选择\n- **建立质量门控**：将评审和验证纳入标准工作流程\n- **匹配任务复杂度**：不要为简单任务过度设计，也不要让复杂任务资源不足\n\n## 🤝 贡献\n\n### 添加新代理\n\n要向集合中贡献一个新的子代理：\n\n1. **遵循命名规范**\n   - 使用小写、连字符分隔的名称（例如 `backend-architect.md`）\n   - 名称应清晰表明代理的领域和角色\n\n2. **使用标准格式**\n   - 包含正确的 frontmatter，包括 `name`、`description` 和可选的 `tools`\n   - 按照 [子代理格式](#-subagent-format) 部分所述的结构化格式进行编写\n\n3. **撰写清晰描述**\n   - 描述应明确指出何时应自动调用该代理\n   - 包括触发代理的具体关键词和情境\n\n4. **定义专业化行为**\n   - 包含详细的任务系统提示，明确角色、专长和能力\n   - 定义与其他代理的交互模式\n   - 规定决策框架和优先级\n\n5. **测试集成**\n   - 验证代理是否能根据描述自动调用\n   - 使用明确请求进行显式调用测试\n   - 确保与现有代理协调模式兼容\n\n### 质量标准\n\n- **领域专长**：代理应在其专业领域展现出深厚的知识\n- **清晰边界**：明确代理负责和不负责的内容\n- **即插即用**：设计为可与其他代理无缝协作\n- **一致语气**：保持专业、友好且具有专家水准的沟通风格\n\n### 提交流程\n\n1. 按照所有标准创建代理文件\n2. 使用各种调用模式测试代理\n3. 提交包含示例用例的拉取请求\n4. 包括关于代理独特价值和集成模式的文档\n\n## 🛠️ 故障排除\n\n**常见问题：**\n\n- **未选择代理**：使用领域特定的关键字或显式调用\n- **意外选择**：提供更多关于技术栈和需求的上下文\n- **通用回复**：要求更具体的深度，并包含详细的约束条件\n- **建议冲突**：请求不同专家之间的协调\n\n**资源：**\n\n- [Claude Code 文档](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) - 官方指南\n- [子代理文档](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code\u002Fsub-agents) - 代理系统参考\n\n## 📊 快速参考\n\n### 最常用的代理\n\n1. **[code-reviewer](agents\u002Fquality-testing\u002Fcode-reviewer.md)** - 质量保证和最佳实践\n2. **[backend-architect](dagents\u002Fevelopment\u002Fbackend-architect.md)** - API 和系统设计\n3. **[frontend-developer](agents\u002Fdevelopment\u002Ffrontend-developer.md)** - UI\u002FUX 实现\n4. **[security-auditor](agents\u002Fsecurity\u002Fsecurity-auditor.md)** - 安全验证和合规\n5. **[performance-engineer](agents\u002Finfrastructure\u002Fperformance-engineer.md)** - 优化和瓶颈分析\n\n### 核心协调模式\n\n- **开发**：`architect → implement → test → review`\n- **调试**：`debugger → specialist → validator`\n- **优化**：`performance-engineer + database-optimizer → validation`\n- **安全**：`primary-agent → security-auditor → approval`\n\n### 成功关键因素\n\n- ✅ 相信自动委派以获得最佳结果\n- ✅ 提供丰富上下文和具体需求\n- ✅ 有策略地使用显式调用\n- ✅ 建立质量门控和验证模式\n- ✅ 从代理协调模式中学习\n\n## 🎬 示例\n\n这些示例展示了真实世界中的多代理协调场景，并附有详细的资源指标，帮助您了解不同项目复杂度下的令牌使用量、执行时间和预期交付成果：\n\n- **示例 1**：简单功能实现（约 30 万令牌，约 17 分钟）——展示针对重点组件开发的高效四代理协调\n- **示例 2**：复杂系统实现（约 85 万令牌，约 45 分钟）——演示企业级七代理编排及错误恢复机制\n\n两个示例均包含实际的令牌计数、执行时间以及交付成果质量，以便您在项目中规划和预算多代理工作流。\n\n### 示例 1：ExportStep 组件实现\n\n**用户请求**：`\u002Fsc:implement` 使用 agent-organizer 设计并实现 ExportStep.tsx 函数，同时优化 UI\u002FUX。\n\n#### 代理编排流程\n\n![代理编排](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_52f11606211e.png)\n\n**步骤 1：agent-organizer 分析**（56.7K 个 token，1 分 20 秒）\n\n- 分析了现有项目结构和 Zustand 状态管理库\n- 制定了包含 7 项任务的完整实现计划\n- 组建了由 3 名专家代理组成的团队以协同执行\n\n![后端实现](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_f75b3e29f83c.png)\n\n**步骤 2：backend-architect 实现**（99.1K 个 token，7 分 31 秒）\n\n- 设计了完整的导出状态管理架构\n- 实现了 SRT、VTT、ASS 和 JSON 格式的格式转换工具\n- 集成了 Electron IPC 以实现无缝文件操作\n\n![前端增强](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_abd528092263.png)\n\n**步骤 3：frontend-developer 增强**（84.3K 个 token，5 分 29 秒）\n\n- 创建了具有真实事件处理程序的完全交互式 ExportStep 组件\n- 实现了基于字幕数据的实时预览生成功能\n- 添加了无障碍合规性（WCAG 2.1 AA）和响应式设计\n\n![测试策略](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_fc90f9d581c7.png)\n\n**步骤 4：test-automator 质量保证**（61.4K 个 token，2 分 46 秒）\n\n- 为格式转换器开发了全面的测试覆盖率\n- 搭建了 Jest 和 React Testing Library 测试框架\n- 制定了无障碍和交互测试策略\n\n#### 实现结果\n\n![最终输出](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_746297a8803a.png)\n\n**完整功能交付**：\n\n- 🏗️ **后端**：带有状态管理、格式转换工具和 Electron IPC 集成的导出状态管理模块\n- 🎨 **前端**：具有实时预览、无障碍合规性和键盘导航功能的交互式 UI\n- ✅ **测试**：具备框架搭建和验证的全面测试覆盖率\n\n![在线演示](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_877b30cc06c5.gif)\n\n#### 项目指标\n\n**资源使用情况**：\n\n- **总 token 数**：约 301K 个 token（agent-organizer：56K，backend-architect：99K，frontend-developer：84K，test-automator：61K）\n- **总耗时**：约 30 分钟执行时间\n- **团队规模**：4 名代理（1 名编排者 + 3 名专家）\n- **创建\u002F修改的文件数**：4 个主要文件（状态管理、组件、工具和测试）\n\n**效率亮点**：\n\n- **顺序协调**：各代理无缝衔接前序工作\n- **高质量集成**：生产就绪的导出系统，功能全面\n- **零破坏性变更**：在不中断现有架构的情况下进行了增强\n\n### 示例 2：复杂的工作区管理系统\n\n**用户请求**：`\u002Fsc:design` 实现一个复杂的工作区管理系统，支持用户配置持久化、多工作区、工作区分组以及类似 Discord 的拖放功能 UI。\n\n#### 第一阶段：全面设计与多代理评估\n\n![代理编排者设计阶段](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_16ba993d2001.png)\n\n**组建 5 人团队**：backend-architect、frontend-developer、electron-pro、ux-designer、test-automator\n\n**设计交付成果**：\n\n- 完整的 TypeScript 接口，涵盖工作区、工作区组及配置\n- IndexedDB 存储策略，并可从 localStorage 迁移\n- 受 Discord 启发的 UI 规范，支持拖放功能\n- 自动保存机制，具备冲突解决和备份策略\n- 包含质量门的 5 阶段实施计划\n\n![第一阶段工作](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_c683a9037368.png)\n\n**第一阶段评估结果**：\n\n![第一阶段完成](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_40e572e12c7c.png)\n![第一阶段总结](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_d472a2472888.png)\n\n**全面团队评估**（5 名代理，总计约 400K 个 token）：\n\n- 🏗️ **后端架构**：IndexedDB 模式、启动时间小于 200 毫秒、迁移框架、自动保存策略\n- 🎨 **前端组件**：受 Discord 启发的设计、Material-UI 集成、渐进增强\n- ⚡ **Electron 集成**：IPC 架构、安全模型、性能优化\n- 🎭 **UX 设计**：A+ UX 评分（92\u002F100），无任何干扰，用户旅程验证\n- ✅ **测试策略**：99.5% 的迁移成功率，4 层测试金字塔，设有质量门\n\n#### 完整实施结果\n\n![所有阶段完成](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_13c183c3e288.png)\n\n**完整的 5 阶段实施**：\n\n- **阶段 1**：评估与现状分析 ✅\n- **阶段 2**：架构定稿与基础设施 ✅\n- **阶段 3**：核心实现 ✅\n- **阶段 4**：集成与迁移 ✅\n- **阶段 5**：质量保证与最终完善 ✅\n\n**最终交付成果**：\n\n- 具有 IndexedDB 持久化的完整工作区管理系统\n- 支持拖放工作的区组织的 Discord 风格 UI\n- 多工作区支持及工作区分组功能\n- 无缝迁移到现有的 localStorage 系统\n- 全面的测试覆盖和错误恢复机制\n\n#### 资源指标与性能\n\n**项目总体指标**：\n\n- **使用的 token 数量**：所有阶段及错误修复共约 900K 个 token\n- **耗时**：总执行时间约 120 分钟\n- **参与的代理**：7 名专业代理（5 名主要代理 + 2 名错误修复代理）\n- **代码行数**：超过 15 个文件中总计约 2,400 行代码\n- **测试覆盖率**：99.5%，涵盖了全面的边界情况（应为幻觉）\n\n#### 嵌套代理协调下的构建错误修复\n\n![构建错误检测](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_aaf81763fef8.png)\n\n**第二次用户指令**：`@agent-code-reviewer-pro` 应用程序存在构建错误，请找出所有构建错误，并请相关子代理进行修复。`@agent-agent-organizer`\n\n![嵌套子代理协调](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_readme_f3d4ad709768.png)\n\n**错误修复流程**：\n\n1. **code-reviewer-pro**（68.5K 个 token，5 分 26 秒）：发现了关键的 TypeScript 语法错误\n2. **agent-organizer** 协调：通过 **typescript-pro** 对构建错误进行系统性修复\n3. **嵌套委派**：在代理工作流中调用专门的子代理进行针对性修复\n\n**错误修复效率**：\n\n- **检测**：code-reviewer-pro 耗时约 5 分钟\n- **协调**：agent-organizer 立即响应\n- **修复实施**：嵌套的 typescript-pro 代理耗时约 30 分钟\n- **构建成功**：经过系统性修复后，不再存在任何错误\n- **挑战性运行时错误**：运行时错误发生，需要手动调试和指导\n\n### 多代理的主要优势：\n\n- **🧠 智能编排**：agent-organizer 协调了 5 名以上代理，完成了复杂的 5 阶段实施\n- **🔧 嵌套代理支持**：通过工作流中的子代理委派进行错误修复\n- **📊 企业级质量**：850K 个 token 的全面分析、设计和实施\n- **⚡ 快速错误恢复**：通过专业代理的协调，在不到 8 分钟内解决了构建错误\n- **🎯 行业专业知识**：每个代理都贡献了其专长领域知识（存储架构、UX 设计、TypeScript 修复）\n\n---\n\n*祝您与 AI 专家团队一起愉快编码！🚀*","# Claude Code Sub-Agents 快速上手指南\n\n`claude-code-sub-agents` 是一个包含 33 个专用 AI 子智能体（Subagents）的集合，旨在通过领域专业知识增强 [Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) 的开发工作流。这些智能体涵盖前端、后端、基础设施、测试、数据\u002FAI 及安全等领域，支持自动委派和多智能体协作。\n\n## 🛠️ 环境准备\n\n在开始之前，请确保满足以下前置条件：\n\n1.  **Claude Code**：已安装并配置好 Anthropic 的 Claude Code CLI 工具。\n2.  **Node.js & npm**：用于安装和运行 MCP 服务器（建议最新 LTS 版本）。\n3.  **Git**：用于克隆仓库或更新代理文件。\n4.  **API Keys**（可选但推荐）：部分高级功能（如 `magic` UI 生成）需要相应的 API Key。\n\n## 📦 安装步骤\n\n### 1. 安装子智能体\n\n推荐使用 Git 克隆方式安装，以便后续轻松更新。智能体文件需放置在 `~\u002F.claude\u002Fagents\u002F` 目录下。\n\n```bash\n# 进入 Claude 配置目录\ncd ~\u002F.claude\n\n# 克隆仓库到 agents 目录\n# 注意：如果 ~\u002F.claude\u002Fagents 已存在且非空，建议先备份或手动复制文件\ngit clone https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents.git agents\n\n# 如果 agents 目录已存在，进入目录拉取最新代码\ncd agents\ngit pull origin main\n```\n\n**手动安装替代方案**（仅复制特定智能体）：\n\n```bash\n# 创建专属目录以避免冲突\nmkdir -p ~\u002F.claude\u002Fagents\u002Flst97\n\n# 将需要的 .md 文件复制到该目录\ncp \u002Fpath\u002Fto\u002Fdownloaded\u002Fagents\u002F*.md ~\u002F.claude\u002Fagents\u002Flst97\n```\n\n### 2. 配置 MCP 服务器（推荐）\n\n为了发挥智能体的最佳性能（如实时文档查询、UI 生成、浏览器自动化等），需要在全球配置文件 `~\u002F.claude.json` 中添加 MCP 服务器配置。\n\n编辑 `~\u002F.claude.json`，添加或合并以下 `mcpServers` 配置：\n\n```json\n{\n  \"mcpServers\": {\n    \"sequential-thinking\": {\n      \"type\": \"stdio\",\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@modelcontextprotocol\u002Fserver-sequential-thinking\"],\n      \"env\": {}\n    },\n    \"context7\": {\n      \"type\": \"stdio\",\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@upstash\u002Fcontext7-mcp\"],\n      \"env\": {}\n    },\n    \"magic\": {\n      \"type\": \"stdio\",\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@21st-dev\u002Fmagic@latest\"],\n      \"env\": {\n        \"API_KEY\": \"your_api_key_here\" \n      }\n    },\n    \"playwright\": {\n      \"type\": \"stdio\",\n      \"command\": \"npx\",\n      \"args\": [\"@playwright\u002Fmcp@latest\"],\n      \"env\": {}\n    },\n    \"filesystem\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@modelcontextprotocol\u002Fserver-filesystem\", \"\u002Fyour\u002Fallowed\u002Fpath\"],\n      \"env\": {}\n    },\n    \"puppeteer\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"puppeteer-mcp-server\"],\n      \"env\": {}\n    }\n  }\n}\n```\n\n> **注意**：\n> - 请将 `\u002Fyour\u002Fallowed\u002Fpath` 替换为你允许文件系统访问的实际路径。\n> - `magic` 服务需要有效的 API Key。\n> - 国内用户若遇到 `npx` 下载缓慢问题，可配置 npm 国内镜像源（如淘宝镜像）加速依赖安装。\n\n### 3. 启用多智能体编排（可选，针对复杂项目）\n\n如果你希望在一个复杂项目中自动协调多个智能体（例如同时调用后端架构师和安全审计员），需在**项目根目录**下启用编排协议。\n\n```bash\n# 将 CLAUDE.md 复制到你的项目根目录\ncp \u002Fpath\u002Fto\u002Fagents\u002FCLAUDE.md \u002Fpath\u002Fto\u002Fyour\u002Fproject\u002FCLAUDE.md\n```\n\n> **警告**：请勿将此 `CLAUDE.md` 放在全局 `~\u002F.claude\u002F` 目录下，否则会导致简单任务也被过度编排，增加 Token 消耗和响应时间。仅在需要复杂协作的大型项目中启用。\n\n## 🚀 基本使用\n\n安装完成后，重启 Claude Code。智能体将根据上下文自动加载。\n\n### 1. 验证安装\n\n在 Claude Code 中输入以下指令，确认智能体已被识别：\n\n```text\nList all available subagents\n```\n\n### 2. 自动调用（推荐）\n\nClaude Code 会根据你的自然语言请求，自动分析上下文并选择最合适的智能体。你无需记住具体的智能体名称。\n\n**示例场景：**\n\n*   **前端开发**：\n    > \"创建一个响应式的 React 登录页面，包含表单验证。\"\n    > *系统可能自动调用：`frontend-developer` 或 `react-pro`*\n\n*   **后端与安全**：\n    > \"实现一个用户注册 API，确保密码加密存储，并进行安全审查。\"\n    > *系统可能自动调用：`backend-architect` -> `security-auditor` -> `test-automator`*\n\n*   **调试与修复**：\n    > \"这个 Python 脚本运行报错，请帮我调试并优化性能。\"\n    > *系统可能自动调用：`python-pro` 和 `debugger`*\n\n### 3. 显式调用\n\n如果你希望指定特定的专家智能体，可以在提示词中明确指明。\n\n**示例：**\n\n```text\nUse the ui-designer agent to suggest a color palette for a fintech app.\n```\n\n```text\nAsk the database-optimizer to review my SQL query for performance issues.\n```\n\n### 4. 多智能体协作（需项目级 CLAUDE.md）\n\n在启用了项目级 `CLAUDE.md` 的复杂项目中，你可以提出跨领域需求，`agent-organizer` 将自动组建团队。\n\n**示例：**\n\n```text\nPlan and implement a new feature for real-time chat, including backend WebSocket setup, frontend React components, and end-to-end tests.\n```\n\n*系统将自动协调 `backend-architect`, `react-pro`, `test-automator` 等智能体协同工作。*","一家初创公司的全栈工程师正面临紧急任务：需要在两天内将现有的单体 Python 后端重构为微服务架构，并同步开发一个高性能的 React 前端仪表盘，同时确保部署流程自动化。\n\n### 没有 claude-code-sub-agents 时\n\n- **上下文切换成本高**：开发者必须在后端逻辑、前端组件和 DevOps 配置之间频繁切换思维模式，导致注意力分散，极易出现低级错误。\n- **专业深度不足**：通用 AI 助手生成的代码往往缺乏特定领域的最佳实践，例如 Python 代码可能未充分利用异步特性，或 React 组件存在不必要的重渲染问题。\n- **架构一致性差**：由于缺乏统一的架构视角，前后端接口定义容易出现偏差，数据库 schema 设计与 API 路由不匹配，导致联调阶段花费大量时间修复集成错误。\n- **部署风险不可控**：手动编写 Dockerfile 和 CI\u002FCD 流水线容易遗漏环境变量或安全配置，导致生产环境部署失败或存在安全隐患。\n\n### 使用 claude-code-sub-agents 后\n\n- **智能自动分工**：claude-code-sub-agents 根据任务上下文自动调用 `backend-architect` 设计微服务边界，`react-pro` 优化前端性能，`deployment-engineer` 配置容器化部署，开发者只需关注核心业务逻辑。\n- **领域专家级代码**：`python-pro` 生成符合 idiomatic 风格的高效后端代码，`frontend-developer` 确保 UI 组件的响应式布局和状态管理符合最佳实践，代码质量显著提升。\n- **端到端协同一致**：`full-stack-developer` 协调前后端接口定义，确保 TypeScript 类型与后端 API 响应严格匹配，大幅减少联调摩擦。\n- **自动化质量保证**：`devops-incident-responder` 预先检查部署配置，`performance-engineer` 识别潜在瓶颈，确保上线即稳定，无需反复试错。\n\n核心价值在于通过领域专用的智能体协作，将复杂的全栈开发任务转化为并行且标准化的工作流，显著提升交付速度与代码可靠性。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flst97_claude-code-sub-agents_5f22c206.png","lst97","SIO TOU LAI [Nelson]","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Flst97_2e0458df.png","Full Stack Developer | Uber Driver 🇭🇰 🇦🇺\r\n",null,"Melbourne, Australia","laisiotou1997@gmail.com","https:\u002F\u002Fwww.lst97.dev","https:\u002F\u002Fgithub.com\u002Flst97",1514,242,"2026-04-02T20:10:21","MIT","Linux, macOS, Windows","未说明",{"notes":92,"python":90,"dependencies":93},"该工具是 Claude Code 的子代理（Subagents）集合，并非独立运行的 AI 模型，因此无本地 GPU\u002F显存需求。核心依赖是安装并配置好 Anthropic 的 Claude Code 环境。需将代理文件复制到 ~\u002F.claude\u002Fagents\u002F 目录。为获得最佳性能，建议在 ~\u002F.claude.json 中配置多个 MCP 服务器（如 sequential-thinking, context7, magic, playwright 等），部分 MCP 服务可能需要 API Key。建议仅在复杂项目中于项目根目录放置 CLAUDE.md 以启用多代理协调，避免全局启用导致简单任务过度编排。",[94,95,96,97,98,99,100,101],"Claude Code CLI","@modelcontextprotocol\u002Fserver-sequential-thinking","@upstash\u002Fcontext7-mcp","@21st-dev\u002Fmagic","@playwright\u002Fmcp","@modelcontextprotocol\u002Fserver-filesystem","puppeteer-mcp-server","npx",[15],[104,105,106,107,108,109],"ai-agents","claude-code","sub-agents","subagents","claudecode-config","claudecode-subagents","2026-03-27T02:49:30.150509","2026-04-06T05:17:52.386305",[113,118,123,128],{"id":114,"question_zh":115,"answer_zh":116,"source_url":117},11628,"如何优化代理（Agents）的文件夹结构以便更好地与 Claude Code 集成？","建议将所有代理文件移动到一个专门的代理文件夹中。如果文件都位于根目录，Claude 可能会对 README.md、CLAUDE.md 等非代理文件产生混淆。通过创建一个单独的代理文件夹，你可以将该文件夹符号链接（symlink）到 .claude 目录，从而简化配置并避免手动逐个链接。维护者已根据此建议更新了项目结构。","https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fissues\u002F8",{"id":119,"question_zh":120,"answer_zh":121,"source_url":122},11629,"是否可以在代理配置中自定义颜色？","虽然理论上可以在每个代理配置的 front matter 中添加 `color:` 字段，但实际上 Claude Code (CC) 会自动为代理分配随机颜色，因此通常无需手动设置颜色字段。","https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fissues\u002F10",{"id":124,"question_zh":125,"answer_zh":126,"source_url":127},11630,"该项目使用什么开源许可证？","维护者已响应用户请求，为项目添加了开源许可证（用户建议 Apache2、BSD 或 MIT，维护者确认将添加）。在使用前请检查仓库根目录下的 LICENSE 文件以确认具体条款。","https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fissues\u002F4",{"id":129,"question_zh":130,"answer_zh":131,"source_url":132},11631,"如何将创建的代理分享到 Subagents 平台？","你可以通过 Subagents 网站直接导入你创建的代理。这样可以保留你的完整署名权，并让更多用户使用。相关讨论和反馈已移至 subagents-feedback 仓库进行。","https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fissues\u002F5",[134,139],{"id":135,"version":136,"summary_zh":137,"released_at":138},62115,"v1.1.0","修复 https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fissues\u002F8 中关于 CC >=1.0.80 的子代理解析问题。\n\n**完整更新日志**：https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fcompare\u002Fv1.0.0...v1.1.0","2025-08-15T11:02:46",{"id":140,"version":141,"summary_zh":142,"released_at":143},62116,"v1.0.0","本集合为 [Claude Code](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) 提供了一套全面的 33 个专用 AI 子代理模板，旨在通过领域-specific 的专业知识和智能自动化来提升开发工作流效率。\n\n## 本次发布的主要特性\n\n*   **33 个专用子代理组成的集合：** 本次发布包含了适用于广泛开发任务的子代理，从前端和后端开发到数据科学和安全性等。\n*   **智能自动委派：** Claude Code 能够根据上下文自动为特定任务选择最优的子代理。\n*   **多代理编排：** `agent-organizer` 子代理可用于管理复杂的多代理任务。\n*   **详尽的文档：** `README.md` 文件提供了项目、其功能以及可用子代理的全面概述。\n*   **多代理协作示例：** `README.md` 文件中包含如何使用这些子代理执行复杂任务的详细示例。\n\n## 贡献\n\n我们欢迎社区的贡献。如果您希望参与该项目，请参阅 `CONTRIBUTING.md` 文件以获取更多信息。\n\n## 变更内容\n* @lachlancooper 在 https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fpull\u002F1 中修复了快速设置说明。\n* @vinod-sc 在 https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fpull\u002F2 中更新了 CLAUDE.md 文件。\n\n## 新贡献者\n* @lachlancooper 在 https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fpull\u002F1 中完成了首次贡献。\n* @vinod-sc 在 https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fpull\u002F2 中完成了首次贡献。\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Flst97\u002Fclaude-code-sub-agents\u002Fcommits\u002Fv1.0.0","2025-08-12T09:13:33"]