[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-rohitg00--awesome-claude-code-toolkit":3,"tool-rohitg00--awesome-claude-code-toolkit":64},[4,16,27,35,48,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":15},8272,"opencode","anomalyco\u002Fopencode","OpenCode 是一款开源的 AI 编程助手（Coding Agent），旨在像一位智能搭档一样融入您的开发流程。它不仅仅是一个代码补全插件，而是一个能够理解项目上下文、自主规划任务并执行复杂编码操作的智能体。无论是生成全新功能、重构现有代码，还是排查难以定位的 Bug，OpenCode 都能通过自然语言交互高效完成，显著减少开发者在重复性劳动和上下文切换上的时间消耗。\n\n这款工具专为软件开发者、工程师及技术研究人员设计，特别适合希望利用大模型能力来提升编码效率、加速原型开发或处理遗留代码维护的专业人群。其核心亮点在于完全开源的架构，这意味着用户可以审查代码逻辑、自定义行为策略，甚至私有化部署以保障数据安全，彻底打破了传统闭源 AI 助手的“黑盒”限制。\n\n在技术体验上，OpenCode 提供了灵活的终端界面（Terminal UI）和正在测试中的桌面应用程序，支持 macOS、Windows 及 Linux 全平台。它兼容多种包管理工具，安装便捷，并能无缝集成到现有的开发环境中。无论您是追求极致控制权的资深极客，还是渴望提升产出的独立开发者，OpenCode 都提供了一个透明、可信",144296,1,"2026-04-16T14:50:03",[13,14],"Agent","插件","ready",{"id":17,"name":18,"github_repo":19,"description_zh":20,"stars":21,"difficulty_score":22,"last_commit_at":23,"category_tags":24,"status":15},6121,"gemini-cli","google-gemini\u002Fgemini-cli","gemini-cli 是一款由谷歌推出的开源 AI 命令行工具，它将强大的 Gemini 大模型能力直接集成到用户的终端环境中。对于习惯在命令行工作的开发者而言，它提供了一条从输入提示词到获取模型响应的最短路径，无需切换窗口即可享受智能辅助。\n\n这款工具主要解决了开发过程中频繁上下文切换的痛点，让用户能在熟悉的终端界面内直接完成代码理解、生成、调试以及自动化运维任务。无论是查询大型代码库、根据草图生成应用，还是执行复杂的 Git 操作，gemini-cli 都能通过自然语言指令高效处理。\n\n它特别适合广大软件工程师、DevOps 人员及技术研究人员使用。其核心亮点包括支持高达 100 万 token 的超长上下文窗口，具备出色的逻辑推理能力；内置 Google 搜索、文件操作及 Shell 命令执行等实用工具；更独特的是，它支持 MCP（模型上下文协议），允许用户灵活扩展自定义集成，连接如图像生成等外部能力。此外，个人谷歌账号即可享受免费的额度支持，且项目基于 Apache 2.0 协议完全开源，是提升终端工作效率的理想助手。",100752,2,"2026-04-10T01:20:03",[14,13,25,26],"图像","开发框架",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":22,"last_commit_at":33,"category_tags":34,"status":15},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[14,26],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":22,"last_commit_at":41,"category_tags":42,"status":15},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。",85092,"2026-04-10T11:13:16",[25,43,44,14,13,45,46,26,47],"数据工具","视频","其他","语言模型","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":15},7525,"codex","openai\u002Fcodex","Codex 是 OpenAI 推出的一款轻量级编程智能体，专为在终端环境中高效运行而设计。它允许开发者直接在命令行界面与 AI 交互，完成代码生成、调试、重构及项目维护等任务，无需频繁切换至浏览器或集成开发环境，从而显著提升了编码流程的连贯性与专注度。\n\n这款工具主要解决了传统 AI 辅助编程中上下文割裂的问题。通过将智能体本地化运行，Codex 能够更紧密地结合当前工作目录的文件结构，提供更具针对性的代码建议，同时支持以自然语言指令驱动复杂的开发操作，让“对话即编码”成为现实。\n\nCodex 非常适合习惯使用命令行的软件工程师、全栈开发者以及技术研究人员。对于追求极致效率、偏好键盘操作胜过图形界面的极客用户而言，它更是理想的结对编程伙伴。\n\n其独特亮点在于灵活的部署方式：既可作为全局命令行工具通过 npm 或 Homebrew 一键安装，也能无缝对接现有的 ChatGPT 订阅计划（如 Plus 或 Pro），直接复用账户权益。此外，它还提供了从纯文本终端到桌面应用的多形态体验，并支持基于 API 密钥的深度定制，充分满足不同场景下的开发需求。",75220,"2026-04-14T14:40:34",[46,13,14],{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":22,"last_commit_at":62,"category_tags":63,"status":15},51,"gstack","garrytan\u002Fgstack","gstack 是 Y Combinator CEO Garry Tan 亲自开源的一套 AI 工程化配置，旨在将 Claude Code 升级为你的虚拟工程团队。面对单人开发难以兼顾产品战略、架构设计、代码审查及质量测试的挑战，gstack 提供了一套标准化解决方案，帮助开发者实现堪比二十人团队的高效产出。\n\n这套配置特别适合希望提升交付效率的创始人、技术负责人，以及初次尝试 Claude Code 的开发者。gstack 的核心亮点在于内置了 15 个具有明确职责的 AI 角色工具，涵盖 CEO、设计师、工程经理、QA 等职能。用户只需通过简单的斜杠命令（如 `\u002Freview` 进行代码审查、`\u002Fqa` 执行测试、`\u002Fplan-ceo-review` 规划功能），即可自动化处理从需求分析到部署上线的全链路任务。\n\n所有操作基于 Markdown 和斜杠命令，无需复杂配置，完全免费且遵循 MIT 协议。gstack 不仅是一套工具集，更是一种现代化的软件工厂实践，让单人开发者也能拥有严谨的工程流程。",73956,"2026-04-16T23:09:21",[13,14],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":80,"owner_email":81,"owner_twitter":82,"owner_website":83,"owner_url":84,"languages":85,"stars":98,"forks":99,"last_commit_at":100,"license":101,"difficulty_score":22,"env_os":102,"env_gpu":103,"env_ram":103,"env_deps":104,"category_tags":112,"github_topics":113,"view_count":22,"oss_zip_url":81,"oss_zip_packed_at":81,"status":15,"created_at":120,"updated_at":121,"faqs":122,"releases":153},8138,"rohitg00\u002Fawesome-claude-code-toolkit","awesome-claude-code-toolkit","The most comprehensive toolkit for Claude Code -- 135 agents, 35 curated skills (+400,000 via SkillKit), 42 commands, 176+ plugins, 20 hooks, 15 rules, 7 templates, 14 MCP configs, 26 companion apps, 52 ecosystem entries, and more.","awesome-claude-code-toolkit 是目前功能最全面的 Claude Code 增强套件，旨在将 Anthropic 的命令行 AI 编程助手打造为强大的自动化开发中心。它通过整合海量社区资源，解决了原生工具在复杂工作流、领域专用能力及生态扩展性上的不足，让用户无需从零配置即可拥有企业级的 AI 协作环境。\n\n这套工具箱特别适合希望大幅提升编码效率的开发者、技术团队负责人以及 AI 工程化研究人员。其核心亮点在于惊人的资源规模：内置 135 个智能体（Agents）处理特定任务，提供 35 项精选技能（并可扩展至 40 万+），包含 176 多个生产级插件、42 条快捷命令及 20 个自动化钩子。此外，它还预置了专业的 MCP 配置、协作规则与模板，甚至集成了如\"CEO\"、“测试工程师”等角色化工作流。无论是进行 AWS 成本优化、代码质量审计，还是构建自主任务队列，awesome-claude-code-toolkit 都能通过一键安装或市场添加，迅速赋予 Claude Code 更深度的上下文理解与执行能力，是进阶用户不可或缺的生态宝库。","# Claude Code Toolkit\n\n**The most comprehensive toolkit for Claude Code -- 135 agents, 35 curated skills (+400,000 via [SkillKit](https:\u002F\u002Fagenstskills.com)), 42 commands, 176+ plugins, 20 hooks, 15 rules, 7 templates, 14 MCP configs, 26 companion apps, 52 ecosystem entries, and more.**\n\n[![Awesome](https:\u002F\u002Fcdn.rawgit.com\u002Fsindresorhus\u002Fawesome\u002Fd7305f38d29fed78fa85652e3a63e154dd8e8829\u002Fmedia\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fsindresorhus\u002Fawesome)\n[![License: Apache-2.0](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache%202.0-blue.svg)](LICENSE)\n[![PRs Welcome](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-welcome-brightgreen.svg)](CONTRIBUTING.md)\n[![Last Updated](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLast%20Updated-Mar%202026-orange.svg)](#)\n[![Files](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FFiles-850+-blueviolet.svg)](#project-structure)\n\n\u003Ca href=\"https:\u002F\u002Fstar-history.com\u002F#rohitg00\u002Fawesome-claude-code-toolkit&Date\">\n  \u003Cpicture>\n    \u003Csource media=\"(prefers-color-scheme: dark)\" srcset=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frohitg00_awesome-claude-code-toolkit_readme_d92a6d46027b.png&theme=dark\" \u002F>\n    \u003Csource media=\"(prefers-color-scheme: light)\" srcset=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frohitg00_awesome-claude-code-toolkit_readme_d92a6d46027b.png\" \u002F>\n    \u003Cimg alt=\"Star History Chart\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frohitg00_awesome-claude-code-toolkit_readme_d92a6d46027b.png\" width=\"600\" \u002F>\n  \u003C\u002Fpicture>\n\u003C\u002Fa>\n\n---\n\n## Quick Install\n\n**Plugin marketplace** (recommended):\n\n```bash\n\u002Fplugin marketplace add rohitg00\u002Fawesome-claude-code-toolkit\n```\n\n**Manual clone:**\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit.git ~\u002F.claude\u002Fplugins\u002Fclaude-code-toolkit\n```\n\n**One-liner:**\n\n```bash\ncurl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fmain\u002Fsetup\u002Finstall.sh | bash\n```\n\n---\n\n## Table of Contents\n\n- [Plugins](#plugins) (176+)\n- [Agents](#agents) (135)\n- [Skills](#skills) (35 curated + 28 community)\n- [Commands](#commands) (42)\n- [Hooks](#hooks) (20 scripts)\n- [Rules](#rules) (15)\n- [Templates](#templates) (7)\n- [MCP Configs](#mcp-configs) (14)\n- [Contexts](#contexts) (5)\n- [Examples](#examples) (3)\n- [Companion Apps & GUIs](#companion-apps--guis)\n- [Ecosystem](#ecosystem)\n- [Setup](#setup)\n- [Resources](#resources)\n- [Related Awesome Lists](#related-awesome-lists)\n- [Contributing](#contributing)\n\n---\n\n## Plugins\n\nOver 176 production-ready plugins that extend Claude Code with domain-specific capabilities.\n\n### Featured\n\n| Plugin | Stars | Description |\n|--------|-------|-------------|\n| [pro-workflow](https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fpro-workflow) | 1,800+ | Battle-tested Claude Code workflows from power users. Self-correcting memory, parallel worktrees, wrap-up rituals, 8 hook types, 5 agents, and the 80\u002F20 AI coding ratio. Install: `\u002Fplugin marketplace add rohitg00\u002Fpro-workflow` |\n| [gstack](https:\u002F\u002Fgithub.com\u002Fgarrytan\u002Fgstack) | 68,200+ | Garry Tan's exact Claude Code setup: 6 opinionated tools that serve as CEO, Eng Manager, Release Manager, and QA Engineer. |\n\n### All Plugins\n\n| Plugin | Description |\n|--------|-------------|\n| [skills-janitor](https:\u002F\u002Fgithub.com\u002Fkhendzel\u002Fskills-janitor) | Audit, deduplicate, check, fix, and track usage of your Claude Code skills. 9 slash commands, zero dependencies |\n| [aws-cost-saver](https:\u002F\u002Fgithub.com\u002Fprajapatimehul\u002Faws-cost-saver) | AWS cost optimization scanner with 173 automated checks, ML-powered rightsizing, and Zero Hallucination Pricing - Real result: 60% cost reduction |\n| [claude-agentic-coding-playbook](https:\u002F\u002Fgithub.com\u002Fjohn-wilmes\u002Fclaude-agentic-coding-playbook) | Evidence-based practices for LLM-assisted development -- hooks, skills, scripts, and a best-practices guide with 58 citations. Includes 19+ guard\u002Flifecycle hooks, investigation workflow, fleet indexing, and claude-loop for autonomous task queues. |\n| [claude-code-mcp](https:\u002F\u002Fgithub.com\u002Fsteipete\u002Fclaude-code-mcp) | Run Claude Code as a one-shot MCP server -- an agent in your agent. Permissions bypassed automatically. By Peter Steinberger. 1,100+ stars |\n| [claude-code-system-prompts](https:\u002F\u002Fgithub.com\u002FPiebald-AI\u002Fclaude-code-system-prompts) | Extracted system prompts from Claude Code -- 18 builtin tool descriptions, sub-agent prompts, utility prompts. Updated for each release. 5,900+ stars |\n| [claude-context](https:\u002F\u002Fgithub.com\u002Fzilliztech\u002Fclaude-context) | Semantic code search MCP server by Zilliz (Milvus creators). Hybrid BM25 + dense vector search. ~40% token reduction. 5,600+ stars |\n| [claude-cost-optimizer](https:\u002F\u002Fgithub.com\u002FSagargupta16\u002Fclaude-cost-optimizer) | Installable cost-mode skill (30-60% cost savings, up to 70% in strict mode) + 10 guides, 10 templates, budget hooks. Install: `npx skills add Sagargupta16\u002Fclaude-cost-optimizer` |\n| [fractal](https:\u002F\u002Fgithub.com\u002Frmolines\u002Ffractal) | Recursive project management for Claude Code. Decomposes goals into predicates, works the riskiest piece first, and re-evaluates as it learns |\n| [a11y-audit](plugins\u002Fa11y-audit\u002F) | Full accessibility audit with WCAG compliance checking |\n| [accessibility-checker](plugins\u002Faccessibility-checker\u002F) | Scan for accessibility issues and fix ARIA attributes in web applications |\n| [adr-writer](plugins\u002Fadr-writer\u002F) | Architecture Decision Records authoring and management |\n| [AgentLint](https:\u002F\u002Fgithub.com\u002F0xmariowu\u002FAgentLint) | Lint your repo for AI agent compatibility. 33 evidence-backed checks across 5 dimensions. Claude Code plugin. |\n| [ai-prompt-lab](plugins\u002Fai-prompt-lab\u002F) | Improve and test AI prompts for better Claude Code interactions |\n| [analytics-reporter](plugins\u002Fanalytics-reporter\u002F) | Generate analytics reports and dashboard configurations from project data |\n| [android-developer](plugins\u002Fandroid-developer\u002F) | Android and Kotlin development with Jetpack Compose |\n| [api-architect](plugins\u002Fapi-architect\u002F) | API design, documentation, and testing with OpenAPI spec generation |\n| [api-benchmarker](plugins\u002Fapi-benchmarker\u002F) | API endpoint benchmarking and performance reporting |\n| [api-reference](plugins\u002Fapi-reference\u002F) | API reference documentation generation from source code |\n| [api-tester](plugins\u002Fapi-tester\u002F) | Test API endpoints and run load tests against services |\n| [aws-helper](plugins\u002Faws-helper\u002F) | AWS service configuration and deployment automation |\n| [azure-helper](plugins\u002Fazure-helper\u002F) | Azure service configuration and deployment automation |\n| [backend-architect](plugins\u002Fbackend-architect\u002F) | Backend service architecture design with endpoint scaffolding |\n| [bug-detective](plugins\u002Fbug-detective\u002F) | Debug issues systematically with root cause analysis and execution tracing |\n| [Bouncer](https:\u002F\u002Fgithub.com\u002Fbuildingopen\u002Fbouncer) | Independent quality gate that uses Gemini to audit Claude Code's output. Includes Stop hook (automatic), quick audit skill, and deep audit with full tool access. One-liner install. |\n| [brooks-lint](https:\u002F\u002Fgithub.com\u002Fhyhmrright\u002Fbrooks-lint) | AI code reviews grounded in six classic engineering books (Brooks, Fowler, Martin, McConnell, Hunt & Thomas, Evans). Diagnoses code across 6 decay risk dimensions with structured findings (Symptom → Source → Consequence → Remedy). Supports PR review, architecture audit, tech debt assessment, and test quality review. v0.6 adds Mermaid dependency graphs. |\n| [ccmanager](https:\u002F\u002Fgithub.com\u002Fkbwo\u002Fccmanager) | Coding agent session manager supporting Claude Code, Gemini CLI, Codex, Cursor, Copilot, Cline, OpenCode, Kimi CLI. Smart auto-approval via Haiku, devcontainer support. 940+ stars |\n| [ccpm](https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fccpm) | Project management using GitHub Issues + Git worktrees for parallel agent execution. Issue-analyze, epic-start, epic-merge commands. 7,600+ stars |\n| [ccusage](https:\u002F\u002Fgithub.com\u002Fryoppippi\u002Fccusage) | CLI for analyzing Claude Code\u002FCodex usage from local JSONL files. Daily, monthly, session, billing-window reports. Offline, zero API calls. 11,500+ stars |\n| [building-multiagent-systems](https:\u002F\u002Fgithub.com\u002F2389-research\u002Fbuilding-multiagent-systems) | Architecture patterns for multi-agent systems: orchestrator-worker, pipeline, debate, and MapReduce topologies |\n| [bundle-analyzer](plugins\u002Fbundle-analyzer\u002F) | Frontend bundle size analysis and tree-shaking optimization |\n| [chief](https:\u002F\u002Fgithub.com\u002FMiniCodeMonkey\u002Fchief) | CLI that wraps Claude Code in a loop. Define a PRD, run chief, go do anything else. Commits after each task, picks up where it left off. Homebrew installable. 380+ stars |\n| [claudebase](https:\u002F\u002Fgithub.com\u002Frohithzr\u002Fclaudebase) | Back up, restore, and sync your Claude Code config (settings, skills, agents, hooks, rules, memory, MCP) to a private GitHub repo. Named profiles, secret scanning, multi-machine conflict detection, automatic backups. 158 tests |\n| [changelog-gen](plugins\u002Fchangelog-gen\u002F) | Generate changelogs from git history with conventional commit parsing |\n| [changelog-writer](plugins\u002Fchangelog-writer\u002F) | Detailed changelog authoring from git history and PRs |\n| [ci-debugger](plugins\u002Fci-debugger\u002F) | Debug CI\u002FCD pipeline failures and fix configurations |\n| [claude-hooks-mastery](https:\u002F\u002Fgithub.com\u002Fdisler\u002Fclaude-code-hooks-mastery) | Complete mastery guide for Claude Code hooks -- UV single-file Python scripts, sub-agents, meta-agent, team-based validation, AI-generated audio feedback. 3,300+ stars |\n| [claude-mem](https:\u002F\u002Fgithub.com\u002Fthedotmack\u002Fclaude-mem) | Automatically captures everything Claude does, compresses with AI, injects relevant context into future sessions. SQLite + full-text search. 35,900+ stars |\n| [claude-notifications-go](https:\u002F\u002Fgithub.com\u002F777genius\u002Fclaude-notifications-go) | Cross-platform smart notifications -- 6 types, click-to-focus, context analysis, webhooks. Single Go binary, zero deps. 340+ stars |\n| [claude-scaffold](https:\u002F\u002Fgithub.com\u002Fpyramidheadshark\u002Fclaude-scaffold) | npx CLI that deploys CLAUDE.md, hooks, and 18 domain skills to any repository in one command. Skills auto-activate via hooks based on project context. Cross-repo sync via `update --all`. Install: `npx claude-scaffold init` |\n| [claude-recap](https:\u002F\u002Fgithub.com\u002Fhatawong\u002Fclaude-recap) | Per-topic session memory using Shell hooks — archives each conversation topic as a separate Markdown summary. Two hooks, bash + Node.js, 100% local |\n| [clirank-mcp-server](https:\u002F\u002Fgithub.com\u002Falexanderclapp\u002Fclirank-mcp-server) ([npm](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002Fclirank-mcp-server)) | API discovery for agents -- scores 210+ APIs on CLI-friendliness across 11 signals. Find the right API for any task via MCP tools or REST (`curl \"https:\u002F\u002Fclirank.dev\u002Fapi\u002Fdiscover?q=send+emails\"`). No auth required |\n| [\u002Fcht](https:\u002F\u002Fgithub.com\u002Fkosoukhov\u002Fcht-cli) | Chat persistence for Claude Code — 13 slash commands + hooks for saving conversations as local markdown files. Crash-proof auto-save via hooks, project-based organization, search, and management. npm: `@kosoukhov\u002Fcht-cli` |\n| [claude-rank](https:\u002F\u002Fgithub.com\u002FHouseofmvps\u002Fclaude-rank) | SEO\u002FGEO\u002FAEO audit plugin — tells you why AI won't cite your site, then auto-fixes robots.txt, sitemap.xml, llms.txt, and JSON-LD. 170+ rules across 10 scanners, zero config |\n| [claude-sounds](https:\u002F\u002Fgithub.com\u002FculminationAI\u002Fclaude-sounds) | Audio feedback for Claude Code hooks — 10 events, 21 sounds, random rotation. macOS (`afplay`). |\n| [claude-code-hooks](https:\u002F\u002Fgithub.com\u002Fyurukusa\u002Fclaude-code-hooks) | 15 production-tested hooks from 160+ hours of autonomous operation. Destructive command blocker, branch guard, syntax check, context monitor, permission auto-approver. Bash, zero deps |\n| [cc-aws-keepalive](https:\u002F\u002Fgithub.com\u002FGeiserX\u002Fcc-aws-keepalive) | Keep Claude Code sessions alive through AWS credential expiry -- proactive expiration warnings, credential refresh bypass for Bedrock SSO\u002FSAML, optional statusline countdown timer. 4 hook scripts, zero dependencies |\n| [cc-safe-setup](https:\u002F\u002Fgithub.com\u002Fyurukusa\u002Fcc-safe-setup) | One command (`npx cc-safe-setup`) to install 6 essential safety hooks in 10 seconds. Zero dependencies |\n| [knowledge-graph](https:\u002F\u002Fgithub.com\u002Fhilyfux\u002Fknowledge-graph) | Built on Anthropic internal engineering practices and Karpathy's AutoResearch methodology. A zero-dependency, git-native memory layer for Claude Code that persists learned context across sessions. Pure bash, ~3ms\u002Fevent, privacy-first |\n| [claude-supermemory](https:\u002F\u002Fgithub.com\u002Fsupermemoryai\u002Fclaude-supermemory) | Persistent memory across sessions and projects using Supermemory. User profile injection at session start, automatic conversation capture. 2,300+ stars |\n| [code-architect](plugins\u002Fcode-architect\u002F) | Generate architecture diagrams and technical design documents |\n| [code-explainer](plugins\u002Fcode-explainer\u002F) | Explain complex code and annotate files with inline documentation |\n| [code-guardian](plugins\u002Fcode-guardian\u002F) | Automated code review, security scanning, and quality enforcement |\n| [code-review-assistant](plugins\u002Fcode-review-assistant\u002F) | Automated code review with severity levels and actionable feedback |\n| [eliniscan](https:\u002F\u002Fgithub.com\u002FAlpYenigun\u002Feliniscan) | AI full codebase scanner. Opens a separate Claude session for every file — reads every line, misses nothing. Finds bugs, security, performance issues and auto-fixes. |\n| [codebase-documenter](plugins\u002Fcodebase-documenter\u002F) | Auto-document entire codebase with inline comments and API docs |\n| [color-contrast](plugins\u002Fcolor-contrast\u002F) | Color contrast checking and accessible color suggestions |\n| [commit-commands](plugins\u002Fcommit-commands\u002F) | Advanced commit workflows with smart staging and push automation |\n| [complexity-reducer](plugins\u002Fcomplexity-reducer\u002F) | Reduce cyclomatic complexity and simplify functions |\n| [compliance-checker](plugins\u002Fcompliance-checker\u002F) | Regulatory compliance verification for GDPR, SOC2, and HIPAA |\n| [content-creator](plugins\u002Fcontent-creator\u002F) | Technical content generation for blog posts and social media |\n| [context7-docs](plugins\u002Fcontext7-docs\u002F) | Fetch up-to-date library documentation via Context7 for accurate coding |\n| [contract-tester](plugins\u002Fcontract-tester\u002F) | API contract testing with Pact for microservice compatibility |\n| [cozempic](https:\u002F\u002Fgithub.com\u002FRuya-AI\u002Fcozempic) | ✨ v1.2.x — Self-updating now, atomic writes, strict session guard, zero false positives on team detection. 13 pruning strategies, Agent Team protection, MCP server, JSONL doctor. Install: `\u002Fplugin marketplace add Ruya-AI\u002Fcozempic` |\n| [create-worktrees](plugins\u002Fcreate-worktrees\u002F) | Git worktree management for parallel development workflows |\n| [cron-scheduler](plugins\u002Fcron-scheduler\u002F) | Cron job configuration and schedule validation |\n| [css-cleaner](plugins\u002Fcss-cleaner\u002F) | Find unused CSS and consolidate stylesheets |\n| [cup](https:\u002F\u002Fgithub.com\u002Fkrodak\u002Fclickup-cli) | ClickUp CLI for AI agents with task management, sprints, and time tracking |\n| [data-privacy](plugins\u002Fdata-privacy\u002F) | Data privacy implementation with PII detection and anonymization |\n| [database-optimizer](plugins\u002Fdatabase-optimizer\u002F) | Database query optimization with index recommendations and EXPLAIN analysis |\n| [dead-code-finder](plugins\u002Fdead-code-finder\u002F) | Find and remove dead code across the codebase |\n| [debug-session](plugins\u002Fdebug-session\u002F) | Interactive debugging workflow with git bisect integration |\n| [dependency-manager](plugins\u002Fdependency-manager\u002F) | Audit, update, and manage project dependencies with safety checks |\n| [deploy-pilot](plugins\u002Fdeploy-pilot\u002F) | Deployment automation with Dockerfile generation, CI\u002FCD pipelines, and infrastructure as code |\n| [desktop-app](plugins\u002Fdesktop-app\u002F) | Desktop application scaffolding with Electron or Tauri |\n| [devops-automator](plugins\u002Fdevops-automator\u002F) | DevOps automation scripts for CI\u002FCD, health checks, and deployments |\n| [discuss](plugins\u002Fdiscuss\u002F) | Debate implementation approaches with structured pros and cons analysis |\n| [claw-army\u002Fclaude-node](https:\u002F\u002Fgithub.com\u002Fclaw-army\u002Fclaude-node) | Python subprocess bridge for Claude Code CLI, giving Python code direct access to Claude Code native capabilities via stream-json |\n| [discoclaw](https:\u002F\u002Fgithub.com\u002FDiscoClaw\u002Fdiscoclaw) | Personal AI orchestrator that bridges Discord to Claude Code with durable memory, task tracking, and cron-based automation |\n| [jarvis](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fjarvis) | Turns an idle Claude Max subscription into a 24\u002F7 AI ops system — Discord bot, 76 scheduled tasks, 12 AI teams, local LanceDB RAG, 98% context compression via Nexus CIG, and 4-layer self-healing infrastructure. Uses `claude -p` headless mode at $0 extra cost. |\n| [jarvis-company-board](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fjarvis-company-board) | Real-time AI agent collaboration board built on Next.js 15 and SQLite WAL — 8 named AI board members debate decisions via SSE push, with a DEV task approval workflow and Railway deploy support. |\n| [dna-claude-analysis](https:\u002F\u002Fgithub.com\u002Fshmlkv\u002Fdna-claude-analysis) | Personal genome analysis toolkit that analyzes raw DNA data across 17 categories and generates a terminal-style HTML dashboard |\n| [codetape](https:\u002F\u002Fgithub.com\u002F888wing\u002Fcodetape) | The flight recorder for AI coding — auto-records semantic traces and syncs README, CHANGELOG, CLAUDE.md. Zero deps. `npx codetape init` |\n| [doc-forge](plugins\u002Fdoc-forge\u002F) | Documentation generation, API docs, and README maintenance |\n| [docker-helper](plugins\u002Fdocker-helper\u002F) | Build optimized Docker images and improve Dockerfile best practices |\n| [double-check](plugins\u002Fdouble-check\u002F) | Verify code correctness with systematic second-pass analysis |\n| [e2e-runner](plugins\u002Fe2e-runner\u002F) | End-to-end test execution and recording for web applications |\n| [embedding-manager](plugins\u002Fembedding-manager\u002F) | Manage vector embeddings and similarity search |\n| [env-manager](plugins\u002Fenv-manager\u002F) | Set up and validate environment configurations across environments |\n| [env-sync](plugins\u002Fenv-sync\u002F) | Environment variable syncing and diff across environments |\n| [experiment-tracker](plugins\u002Fexperiment-tracker\u002F) | ML experiment tracking with metrics logging and run comparison |\n| [explore](plugins\u002Fexplore\u002F) | Smart codebase exploration with dependency mapping and structure analysis |\n| [faf-skills](https:\u002F\u002Fgithub.com\u002FWolfe-Jam\u002Ffaf-skills) | 31 Claude Code skills for persistent project context (.faf, `application\u002Fvnd.faf+yaml`). Scoring, sync, testing, publishing, MCP server creation, architecture docs. Anthropic MCP #2759. |\n| [feature-dev](plugins\u002Ffeature-dev\u002F) | Full feature development workflow from spec to completion |\n| [fractal](https:\u002F\u002Fgithub.com\u002Frmolines\u002Ffractal) | Recursive project management plugin. Decomposes any goal into verifiable predicates, works on the riskiest unknown first. Features `\u002Ffractal:run` (idempotent state machine), `\u002Ffractal:init`, `\u002Ffractal:patch`, dry run mode, and incremental decomposition with re-evaluation. |\n| [finance-tracker](plugins\u002Ffinance-tracker\u002F) | Development cost tracking with time estimates and budget reporting |\n| [fix-github-issue](plugins\u002Ffix-github-issue\u002F) | Auto-fix GitHub issues by analyzing issue details and implementing solutions |\n| [fix-pr](plugins\u002Ffix-pr\u002F) | Fix PR review comments automatically with context-aware patches |\n| [flutter-mobile](plugins\u002Fflutter-mobile\u002F) | Flutter app development with widget creation and platform channels |\n| [frontend-developer](plugins\u002Ffrontend-developer\u002F) | Frontend component development with accessibility and responsive design |\n| [gcp-helper](plugins\u002Fgcp-helper\u002F) | Google Cloud Platform service configuration and deployment |\n| [git-flow](plugins\u002Fgit-flow\u002F) | Git workflow management with feature branches, releases, and hotfix flows |\n| [github-issue-manager](plugins\u002Fgithub-issue-manager\u002F) | GitHub issue triage, creation, and management |\n| [harness-evolver](https:\u002F\u002Fgithub.com\u002Fraphaelchristi\u002Fharness-evolver) | LangSmith-native autonomous agent optimization. Multi-agent proposers evolve prompts, routing, tools, and architecture in isolated git worktrees. Install: `npx harness-evolver@latest` |\n| [helm-charts](plugins\u002Fhelm-charts\u002F) | Helm chart generation and upgrade management |\n| [import-organizer](plugins\u002Fimport-organizer\u002F) | Organize, sort, and clean import statements |\n| [infrastructure-maintainer](plugins\u002Finfrastructure-maintainer\u002F) | Infrastructure maintenance with security audits and update management |\n| [ios-developer](plugins\u002Fios-developer\u002F) | iOS and Swift development with SwiftUI views and models |\n| [k8s-helper](plugins\u002Fk8s-helper\u002F) | Generate Kubernetes manifests and debug pod issues with kubectl |\n| [leapfrog-mcp](https:\u002F\u002Fgithub.com\u002Fanthonybono21-cloud\u002Fleapfrog) | Multi-session browser MCP server -- 15 parallel isolated Chromium sessions so multiple Claude Code terminals can browse simultaneously. 37 tools for navigation, extraction, interaction, network interception, and session profiles. 797 tests. Install: `npx -y leapfrog-mcp` |\n| [license-checker](plugins\u002Flicense-checker\u002F) | License compliance checking and NOTICE file generation |\n| [lighthouse-runner](plugins\u002Flighthouse-runner\u002F) | Run Lighthouse audits and fix performance issues |\n| [lightcms](https:\u002F\u002Fgithub.com\u002Fjonradoff\u002Flightcms) | AI-native CMS with 41 MCP tools for managing websites through natural language — pages, templates, assets, themes, collections, redirects, and more with full content versioning |\n| [linear-helper](plugins\u002Flinear-helper\u002F) | Linear issue tracking integration and workflow management |\n| [load-tester](plugins\u002Fload-tester\u002F) | Load and stress testing for APIs and web services |\n| [memory-profiler](plugins\u002Fmemory-profiler\u002F) | Memory leak detection and heap analysis |\n| [migrate-tool](plugins\u002Fmigrate-tool\u002F) | Generate database migrations and code migration scripts for framework upgrades |\n| [migration-generator](plugins\u002Fmigration-generator\u002F) | Database migration generation and rollback management |\n| [model-context-protocol](plugins\u002Fmodel-context-protocol\u002F) | MCP server development helper with tool and resource scaffolding |\n| [model-evaluator](plugins\u002Fmodel-evaluator\u002F) | Evaluate and compare ML model performance metrics |\n| [monitoring-setup](plugins\u002Fmonitoring-setup\u002F) | Monitoring and alerting configuration with dashboard generation |\n| [monorepo-manager](plugins\u002Fmonorepo-manager\u002F) | Manage monorepo packages with affected detection and version synchronization |\n| [mutation-tester](plugins\u002Fmutation-tester\u002F) | Mutation testing to measure test suite quality |\n| [myclaude](https:\u002F\u002Fgithub.com\u002Fstellarlinkco\u002Fmyclaude) | Multi-agent orchestration routing tasks to Claude Code, Codex, Gemini, and OpenCode based on complexity. OmO skill for intelligent routing. 2,400+ stars |\n| [nimbalyst](https:\u002F\u002Fnimbalyst.com) | Visual workspace for building with Codex and Claude Code. Session and task manager. Visual editing |\n| [n8n-workflow](plugins\u002Fn8n-workflow\u002F) | Generate n8n automation workflows from natural language descriptions |\n| [onboarding-guide](plugins\u002Fonboarding-guide\u002F) | New developer onboarding documentation generator |\n| [oh-my-claudecode](https:\u002F\u002Fgithub.com\u002FYeachan-Heo\u002Foh-my-claudecode) | Teams-first multi-agent orchestration. 19 specialized agents, 28 skills. Full autonomous execution, Socratic questioning, N coordinated agents. 9,900+ stars |\n| [onWatch](https:\u002F\u002Fgithub.com\u002Fonllm-dev\u002Fonwatch) | Open-source Go CLI that tracks AI API quota usage across 7 providers (Synthetic, Z.ai, Anthropic, Codex, GitHub Copilot, MiniMax, Antigravity) with a background daemon (\u003C50 MB RAM), zero telemetry, and a Material Design 3 web dashboard |\n| [obey](https:\u002F\u002Fgithub.com\u002FLexxes-Projects\u002Fobey) | Rule enforcement plugin. Save rules with natural language, enforce with 17 lifecycle hooks. Three scopes (global, stack-specific, project-local), active blocking via PreToolUse, completion checklists via Stop hook, audit trail, auto-detects rule-like language. Cross-platform. |\n| [opcode](https:\u002F\u002Fgithub.com\u002Fwinfunc\u002Fopcode) | Tauri 2 desktop GUI and toolkit for Claude Code. Manage sessions, create custom agents with visual editor, usage analytics, MCP integration. 21,000+ stars |\n| [openapi-expert](plugins\u002Fopenapi-expert\u002F) | OpenAPI spec generation, validation, and client code scaffolding |\n| [optimize](plugins\u002Foptimize\u002F) | Code optimization for performance and bundle size reduction |\n| [ORCH](https:\u002F\u002Fgithub.com\u002Foxgeneral\u002FORCH) | CLI runtime orchestrating Claude Code, Codex, and Cursor as typed agent teams with state machine (todo→review→done), auto-retry, inter-agent messaging, goals, and TUI dashboard |\n| [oss-autopilot](https:\u002F\u002Fgithub.com\u002Fcostajohnt\u002Foss-autopilot) | Open source contribution manager — tracks PRs across repos, discovers issues, diagnoses CI failures, drafts maintainer responses. 7 agents, interactive commands, MCP server. Install: `\u002Fplugin marketplace add costajohnt\u002Foss-autopilot` |\n| [paco-framework](https:\u002F\u002Fgithub.com\u002FPenguinAlleyApps\u002Fpaco-framework) | 1+ | Markdown-first multi-agent OS for Claude Code. Coordinates 3-16 specialized AI agents (Engineering, QA, Growth, Finance) with file-based dispatch, institutional memory, CEO Gate approval, and 24\u002F7 scheduling. No Python required. MIT |\n| [peon-ping](https:\u002F\u002Fgithub.com\u002FPeonPing\u002Fpeon-ping) | Warcraft III Peon voice notifications (+ StarCraft, Portal, Zelda) for Claude Code and other agents. Desktop banners, auto-detects SSH\u002Fdevcontainers. 3,900+ stars |\n| [perf-profiler](plugins\u002Fperf-profiler\u002F) | Performance analysis, profiling, and optimization recommendations |\n| [performance-monitor](plugins\u002Fperformance-monitor\u002F) | Profile API endpoints and run benchmarks to identify performance bottlenecks |\n| [plan](plugins\u002Fplan\u002F) | Structured planning with risk assessment and time estimation |\n| [preflight](https:\u002F\u002Fgithub.com\u002Fpreflight-dev\u002Fpreflight) | 24-tool MCP server that catches vague prompts before they cost 2-3x in wrong→fix cycles. 12-category scorecards, correction pattern learning, cross-service contract awareness, session history search with LanceDB vectors, and cost estimation. `npx preflight-dev` |\n| [pr-reviewer](plugins\u002Fpr-reviewer\u002F) | Review pull requests with structured analysis and approve with confidence |\n| [product-shipper](plugins\u002Fproduct-shipper\u002F) | Ship features end-to-end with launch checklists and rollout plans |\n| [production-grade](https:\u002F\u002Fgithub.com\u002Fnagisanzenin\u002Fclaude-code-production-grade-plugin) | 14-agent autonomous pipeline — PM, Architect, Backend, Frontend, QA, Security, Code Review, DevOps, SRE, Data Scientist, Technical Writer, Skill Maker, Polymath co-pilot. Two-wave parallel execution, brownfield-safe. |\n| [project-scaffold](plugins\u002Fproject-scaffold\u002F) | Scaffold new projects and add features with best-practice templates |\n| [prompt-optimizer](plugins\u002Fprompt-optimizer\u002F) | Analyze and optimize AI prompts for better results |\n| [pulse](https:\u002F\u002Fgithub.com\u002Fchsm04\u002Fpulse) | Local Channel plugin — push notifications into Claude Code sessions via HTTP POST. No Discord\u002FSlack needed, just curl. Three levels (info\u002Fwarn\u002Ferror), source tracking, dedup. |\n| [PUIUX Pilot](https:\u002F\u002Fgithub.com\u002FPUIUX-Cloud\u002Fpuiux-pilot) | Auto-configures Claude Code hooks, MCPs, and skills for any project. Scans 95+ project types, selects from 28+ hooks, scores quality (0-100), translates configs across AI tools. Safe: dry-run, atomic writes, backup + rollback. |\n| [python-expert](plugins\u002Fpython-expert\u002F) | Python-specific development with type hints and idiomatic refactoring |\n| [pulser](https:\u002F\u002Fgithub.com\u002Fwhynowlab\u002Fpulser) | Diagnostic CLI for Claude Code SKILL.md files — scans 8 rules, auto-classifies, prescribes fixes with templates, --fix with rollback |\n| [query-optimizer](plugins\u002Fquery-optimizer\u002F) | SQL query optimization and execution plan analysis |\n| [rag-builder](plugins\u002Frag-builder\u002F) | Build Retrieval-Augmented Generation pipelines |\n| [rapid-prototyper](plugins\u002Frapid-prototyper\u002F) | Quick prototype scaffolding with minimal viable structure |\n| [reporecall](https:\u002F\u002Fgithub.com\u002Fproofofwork-agency\u002Freporecall) | Local codebase memory for Claude Code. Tree-sitter AST indexing (22 languages), hybrid keyword + vector search, call-graph traversal, hooks + MCP. Injects relevant context in ~5ms before Claude starts thinking. npm: `@proofofwork-agency\u002Freporecall` |\n| [react-native-dev](plugins\u002Freact-native-dev\u002F) | React Native mobile development with platform-specific optimizations |\n| [readme-generator](plugins\u002Freadme-generator\u002F) | Smart README generation from project analysis |\n| [refactor-engine](plugins\u002Frefactor-engine\u002F) | Extract functions, simplify complex code, and reduce cognitive complexity |\n| [regex-builder](plugins\u002Fregex-builder\u002F) | Build, test, and debug regular expression patterns |\n| [release-manager](plugins\u002Frelease-manager\u002F) | Semantic versioning management and automated release workflows |\n| [responsive-designer](plugins\u002Fresponsive-designer\u002F) | Responsive design implementation and testing |\n| [review-squad](https:\u002F\u002Fgithub.com\u002F2389-research\u002Freview-squad) | Multi-perspective code review via dispatched subagent panels: experts, normie users, pedantic nitpickers, and real-user task runners |\n| [schema-designer](plugins\u002Fschema-designer\u002F) | Database schema design and ERD generation |\n| [screen-reader-tester](plugins\u002Fscreen-reader-tester\u002F) | Screen reader compatibility testing and ARIA fixes |\n| [security-guidance](plugins\u002Fsecurity-guidance\u002F) | Security best practices advisor with vulnerability detection and fixes |\n| [seed-generator](plugins\u002Fseed-generator\u002F) | Database seeding script generation with realistic data |\n| [sitemd](https:\u002F\u002Fgithub.com\u002Fsitemd-cc\u002Fsitemd) | Build websites from Markdown via MCP. 22 tools for creating pages, generating content, configuring settings, running SEO audits, and deploying static sites to Cloudflare Pages |\n| [simmer](https:\u002F\u002Fgithub.com\u002F2389-research\u002Fsimmer) | Iterative artifact refinement using judge subagents that score against user-defined criteria across multiple rounds |\n| [slack-notifier](plugins\u002Fslack-notifier\u002F) | Slack integration for deployment and build notifications |\n| [smart-commit](plugins\u002Fsmart-commit\u002F) | Intelligent git commits with conventional format, semantic analysis, and changelog generation |\n| [sprint-prioritizer](plugins\u002Fsprint-prioritizer\u002F) | Sprint planning with story prioritization and capacity estimation |\n| [technical-sales](plugins\u002Ftechnical-sales\u002F) | Technical demo creation and POC proposal writing |\n| [the-pragmatic-pm](https:\u002F\u002Fgithub.com\u002Fmarfoerst\u002Fthe-pragmatic-pm) | PM leadership toolkit with 43 skills, 5 agents, 4 workflows. Covers PRD generation, OKR lifecycle, pricing, AI pricing, positioning, sales enablement, and quarterly planning. |\n| [terraform-helper](plugins\u002Fterraform-helper\u002F) | Terraform module creation and infrastructure planning |\n| [test-data-generator](plugins\u002Ftest-data-generator\u002F) | Generate realistic test data and seed databases |\n| [test-kitchen](https:\u002F\u002Fgithub.com\u002F2389-research\u002Ftest-kitchen) | Parallel implementation exploration using competing subagents, with structured comparison and winner selection |\n| [test-results-analyzer](plugins\u002Ftest-results-analyzer\u002F) | Analyze test failures, identify patterns, and suggest targeted fixes |\n| [test-writer](plugins\u002Ftest-writer\u002F) | Generate comprehensive unit and integration tests with full coverage |\n| [tool-evaluator](plugins\u002Ftool-evaluator\u002F) | Evaluate and compare developer tools with structured scoring criteria |\n| [type-migrator](plugins\u002Ftype-migrator\u002F) | Migrate JavaScript files to TypeScript with proper types |\n| [ui-designer](plugins\u002Fui-designer\u002F) | Implement UI designs from specs with pixel-perfect component generation |\n| [ultrathink](plugins\u002Fultrathink\u002F) | Deep analysis mode with extended reasoning for complex problems |\n| [unit-test-generator](plugins\u002Funit-test-generator\u002F) | Generate comprehensive unit tests for any function or module |\n| [update-branch](plugins\u002Fupdate-branch\u002F) | Rebase and update feature branches with conflict resolution |\n| [vision-specialist](plugins\u002Fvision-specialist\u002F) | Image and visual analysis with screenshot interpretation and text extraction |\n| [vibe-kanban](https:\u002F\u002Fgithub.com\u002FBloopAI\u002Fvibe-kanban) | Kanban-based orchestration for 10+ coding agents (Claude Code, Codex, Gemini CLI, Copilot, Amp). Isolated git worktrees per agent, inline diff review. 23,200+ stars |\n| [VibeGuard](https:\u002F\u002Fgithub.com\u002Fmajiayu000\u002Fvibeguard) | Stop AI from hallucinating code. 88 rules + 13 hooks + 14 agents for real-time interception and static scanning across 5 languages |\n| [visual-regression](plugins\u002Fvisual-regression\u002F) | Visual regression testing with screenshot comparison |\n| [wshobson\u002Fagents](https:\u002F\u002Fgithub.com\u002Fwshobson\u002Fagents) | 112 specialized agents, 16 multi-agent workflow orchestrators, 146 skills, 79 tools in 72 focused plugins. 31,300+ stars |\n| [web-dev](plugins\u002Fweb-dev\u002F) | Full-stack web development with app scaffolding and page generation |\n| [workflow-optimizer](plugins\u002Fworkflow-optimizer\u002F) | Development workflow analysis and optimization recommendations |\n| [background-timer](https:\u002F\u002Fgithub.com\u002FculminationAI\u002Fbackground-timer) | Background timer with task notifications -- set delayed checks without blocking conversation |\n| [claude-sounds](https:\u002F\u002Fgithub.com\u002FculminationAI\u002Fclaude-sounds) | Audio feedback for Claude Code hooks -- 10 events, 21 sounds, random rotation, customizable (macOS) |\n\n| [notch-so-good](https:\u002F\u002Fgithub.com\u002Fdeepshal99\u002Fnotch-so-good) | Pixel-art crab (Chawd) lives in your Mac's notch and watches Claude Code for you. Live session timers, color-coded notifications, 13 idle animations, mouse-reactive eyes, drowsiness system. Universal binary, one-line install: `npx notch-so-good`. MIT, 130+ users |\n| [codebase-graph](https:\u002F\u002Fgithub.com\u002FPhoenixrr2113\u002Fcodebase-graph) | Code intelligence MCP server — 42-language tree-sitter AST parsing, FalkorDB knowledge graphs, 0.944 MRR search quality. npm: `@anthropic\u002Fcodegraph` |\n\n### Installing a Plugin\n\n```bash\n\u002Fplugin install claude-code-toolkit@smart-commit\n```\n\nOr install all plugins at once:\n\n```bash\n\u002Fplugin install claude-code-toolkit\n```\n\n---\n\n## Agents\n\nOne hundred thirty-five specialized agents organized into ten categories. Each agent defines a persona, system instructions, and tool access patterns.\n\n### Core Development (13 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| Fullstack Engineer | [`fullstack-engineer.md`](agents\u002Fcore-development\u002Ffullstack-engineer.md) | End-to-end feature delivery across frontend, backend, and database |\n| API Designer | [`api-designer.md`](agents\u002Fcore-development\u002Fapi-designer.md) | RESTful API design with OpenAPI, versioning, and pagination |\n| Frontend Architect | [`frontend-architect.md`](agents\u002Fcore-development\u002Ffrontend-architect.md) | Component architecture, state management, performance |\n| Mobile Developer | [`mobile-developer.md`](agents\u002Fcore-development\u002Fmobile-developer.md) | Cross-platform mobile with React Native and Flutter |\n| Backend Developer | [`backend-developer.md`](agents\u002Fcore-development\u002Fbackend-developer.md) | Node.js\u002FExpress\u002FFastify backend services |\n| GraphQL Architect | [`graphql-architect.md`](agents\u002Fcore-development\u002Fgraphql-architect.md) | Schema design, resolvers, federation, DataLoader |\n| Microservices Architect | [`microservices-architect.md`](agents\u002Fcore-development\u002Fmicroservices-architect.md) | Distributed systems, event-driven, saga patterns |\n| WebSocket Engineer | [`websocket-engineer.md`](agents\u002Fcore-development\u002Fwebsocket-engineer.md) | Real-time communication, Socket.io, scaling |\n| UI Designer | [`ui-designer.md`](agents\u002Fcore-development\u002Fui-designer.md) | UI\u002FUX implementation, design systems, Figma-to-code |\n| Electron Developer | [`electron-developer.md`](agents\u002Fcore-development\u002Felectron-developer.md) | Electron desktop apps, IPC, native OS integration |\n| API Gateway Engineer | [`api-gateway-engineer.md`](agents\u002Fcore-development\u002Fapi-gateway-engineer.md) | API gateway patterns, rate limiting, auth proxies |\n| Monorepo Architect | [`monorepo-architect.md`](agents\u002Fcore-development\u002Fmonorepo-architect.md) | Turborepo\u002FNx workspace strategies, dependency graphs |\n| Event-Driven Architect | [`event-driven-architect.md`](agents\u002Fcore-development\u002Fevent-driven-architect.md) | Event sourcing, CQRS, message queues, distributed events |\n\n### Language Experts (25 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| TypeScript | [`typescript-specialist.md`](agents\u002Flanguage-experts\u002Ftypescript-specialist.md) | Type-safe patterns, generics, module design |\n| Python | [`python-engineer.md`](agents\u002Flanguage-experts\u002Fpython-engineer.md) | Pythonic patterns, packaging, async |\n| Rust | [`rust-systems.md`](agents\u002Flanguage-experts\u002Frust-systems.md) | Ownership, lifetimes, trait design |\n| Go | [`golang-developer.md`](agents\u002Flanguage-experts\u002Fgolang-developer.md) | Interfaces, goroutines, error handling |\n| Next.js | [`nextjs-developer.md`](agents\u002Flanguage-experts\u002Fnextjs-developer.md) | App Router, RSC, ISR, server actions |\n| React | [`react-specialist.md`](agents\u002Flanguage-experts\u002Freact-specialist.md) | React 19, hooks, state management |\n| Django | [`django-developer.md`](agents\u002Flanguage-experts\u002Fdjango-developer.md) | Django 5+, DRF, ORM optimization |\n| Rails | [`rails-expert.md`](agents\u002Flanguage-experts\u002Frails-expert.md) | Rails 7+, Hotwire, ActiveRecord |\n| Java | [`java-architect.md`](agents\u002Flanguage-experts\u002Fjava-architect.md) | Spring Boot 3+, JPA, microservices |\n| Kotlin | [`kotlin-specialist.md`](agents\u002Flanguage-experts\u002Fkotlin-specialist.md) | Coroutines, Ktor, multiplatform |\n| Flutter | [`flutter-expert.md`](agents\u002Flanguage-experts\u002Fflutter-expert.md) | Flutter 3+, Dart, Riverpod |\n| C# | [`csharp-developer.md`](agents\u002Flanguage-experts\u002Fcsharp-developer.md) | .NET 8+, ASP.NET Core, EF Core |\n| PHP | [`php-developer.md`](agents\u002Flanguage-experts\u002Fphp-developer.md) | PHP 8.3+, Laravel 11, Eloquent |\n| Elixir | [`elixir-expert.md`](agents\u002Flanguage-experts\u002Felixir-expert.md) | OTP, Phoenix LiveView, Ecto |\n| Angular | [`angular-architect.md`](agents\u002Flanguage-experts\u002Fangular-architect.md) | Angular 17+, signals, standalone components |\n| Vue | [`vue-specialist.md`](agents\u002Flanguage-experts\u002Fvue-specialist.md) | Vue 3, Composition API, Pinia, Nuxt |\n| Svelte | [`svelte-developer.md`](agents\u002Flanguage-experts\u002Fsvelte-developer.md) | SvelteKit, runes, form actions |\n| Swift | [`swift-developer.md`](agents\u002Flanguage-experts\u002Fswift-developer.md) | SwiftUI, iOS 17+, Combine, structured concurrency |\n| Scala | [`scala-developer.md`](agents\u002Flanguage-experts\u002Fscala-developer.md) | Akka actors, Play Framework, Cats Effect |\n| Haskell | [`haskell-developer.md`](agents\u002Flanguage-experts\u002Fhaskell-developer.md) | Pure FP, monads, type classes, GHC extensions |\n| Lua | [`lua-developer.md`](agents\u002Flanguage-experts\u002Flua-developer.md) | Game scripting, Neovim plugins, LuaJIT |\n| Zig | [`zig-developer.md`](agents\u002Flanguage-experts\u002Fzig-developer.md) | Systems programming, comptime, allocator strategies |\n| Clojure | [`clojure-developer.md`](agents\u002Flanguage-experts\u002Fclojure-developer.md) | REPL-driven development, Ring\u002FCompojure, ClojureScript |\n| OCaml | [`ocaml-developer.md`](agents\u002Flanguage-experts\u002Focaml-developer.md) | Type inference, pattern matching, Dream framework |\n| Nim | [`nim-developer.md`](agents\u002Flanguage-experts\u002Fnim-developer.md) | Metaprogramming, GC strategies, C\u002FC++ interop |\n\n### Infrastructure (11 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| Cloud Architect | [`cloud-architect.md`](agents\u002Finfrastructure\u002Fcloud-architect.md) | AWS, GCP, Azure provisioning and IaC |\n| DevOps Engineer | [`devops-engineer.md`](agents\u002Finfrastructure\u002Fdevops-engineer.md) | CI\u002FCD, containerization, monitoring |\n| Database Admin | [`database-admin.md`](agents\u002Finfrastructure\u002Fdatabase-admin.md) | Schema design, query tuning, replication |\n| Platform Engineer | [`platform-engineer.md`](agents\u002Finfrastructure\u002Fplatform-engineer.md) | Internal developer platforms, service catalogs |\n| Kubernetes Specialist | [`kubernetes-specialist.md`](agents\u002Finfrastructure\u002Fkubernetes-specialist.md) | Operators, CRDs, service mesh, Istio |\n| Terraform Engineer | [`terraform-engineer.md`](agents\u002Finfrastructure\u002Fterraform-engineer.md) | IaC, module design, state management, multi-cloud |\n| Network Engineer | [`network-engineer.md`](agents\u002Finfrastructure\u002Fnetwork-engineer.md) | DNS, load balancers, CDN, firewall rules |\n| SRE Engineer | [`sre-engineer.md`](agents\u002Finfrastructure\u002Fsre-engineer.md) | SLOs, error budgets, incident response, postmortems |\n| Deployment Engineer | [`deployment-engineer.md`](agents\u002Finfrastructure\u002Fdeployment-engineer.md) | Blue-green, canary releases, rolling updates |\n| Security Engineer | [`security-engineer.md`](agents\u002Finfrastructure\u002Fsecurity-engineer.md) | IAM policies, mTLS, secrets management, Vault |\n| Incident Responder | [`incident-responder.md`](agents\u002Finfrastructure\u002Fincident-responder.md) | Incident triage, runbooks, communication, recovery |\n\n### Quality Assurance (10 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| Code Reviewer | [`code-reviewer.md`](agents\u002Fquality-assurance\u002Fcode-reviewer.md) | PR review with security and performance focus |\n| Test Architect | [`test-architect.md`](agents\u002Fquality-assurance\u002Ftest-architect.md) | Test strategy, pyramid, coverage targets |\n| Security Auditor | [`security-auditor.md`](agents\u002Fquality-assurance\u002Fsecurity-auditor.md) | Vulnerability scanning, OWASP compliance |\n| Performance Engineer | [`performance-engineer.md`](agents\u002Fquality-assurance\u002Fperformance-engineer.md) | Load testing, profiling, optimization |\n| Accessibility Specialist | [`accessibility-specialist.md`](agents\u002Fquality-assurance\u002Faccessibility-specialist.md) | WCAG compliance, ARIA, screen readers |\n| Chaos Engineer | [`chaos-engineer.md`](agents\u002Fquality-assurance\u002Fchaos-engineer.md) | Chaos testing, fault injection, resilience validation |\n| Penetration Tester | [`penetration-tester.md`](agents\u002Fquality-assurance\u002Fpenetration-tester.md) | OWASP Top 10 assessment, vulnerability reporting |\n| QA Automation | [`qa-automation.md`](agents\u002Fquality-assurance\u002Fqa-automation.md) | Test automation frameworks, CI integration |\n| Compliance Auditor | [`compliance-auditor.md`](agents\u002Fquality-assurance\u002Fcompliance-auditor.md) | SOC 2, GDPR, HIPAA compliance checking |\n| Error Detective | [`error-detective.md`](agents\u002Fquality-assurance\u002Ferror-detective.md) | Error tracking, stack trace analysis, root cause ID |\n\n### Data & AI (15 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| AI Engineer | [`ai-engineer.md`](agents\u002Fdata-ai\u002Fai-engineer.md) | AI application integration, RAG, agents |\n| ML Engineer | [`ml-engineer.md`](agents\u002Fdata-ai\u002Fml-engineer.md) | ML pipelines, training, evaluation |\n| Data Scientist | [`data-scientist.md`](agents\u002Fdata-ai\u002Fdata-scientist.md) | Statistical analysis, visualization |\n| Data Engineer | [`data-engineer.md`](agents\u002Fdata-ai\u002Fdata-engineer.md) | ETL pipelines, Spark, data warehousing |\n| LLM Architect | [`llm-architect.md`](agents\u002Fdata-ai\u002Fllm-architect.md) | Fine-tuning, model selection, serving |\n| Prompt Engineer | [`prompt-engineer.md`](agents\u002Fdata-ai\u002Fprompt-engineer.md) | Prompt optimization, structured outputs |\n| MLOps Engineer | [`mlops-engineer.md`](agents\u002Fdata-ai\u002Fmlops-engineer.md) | Model serving, monitoring, A\u002FB testing |\n| NLP Engineer | [`nlp-engineer.md`](agents\u002Fdata-ai\u002Fnlp-engineer.md) | NLP pipelines, embeddings, classification |\n| Database Optimizer | [`database-optimizer.md`](agents\u002Fdata-ai\u002Fdatabase-optimizer.md) | Query optimization, indexing, partitioning |\n| Computer Vision | [`computer-vision-engineer.md`](agents\u002Fdata-ai\u002Fcomputer-vision-engineer.md) | Image classification, object detection, PyTorch |\n| Recommendation Engine | [`recommendation-engine.md`](agents\u002Fdata-ai\u002Frecommendation-engine.md) | Collaborative filtering, content-based, hybrid |\n| ETL Specialist | [`etl-specialist.md`](agents\u002Fdata-ai\u002Fetl-specialist.md) | Data pipelines, schema evolution, data quality |\n| Vector DB Engineer | [`vector-database-engineer.md`](agents\u002Fdata-ai\u002Fvector-database-engineer.md) | FAISS, Pinecone, Qdrant, Weaviate, embeddings |\n| Data Visualization | [`data-visualization.md`](agents\u002Fdata-ai\u002Fdata-visualization.md) | D3.js, Chart.js, Matplotlib, Plotly dashboards |\n| Feature Engineer | [`feature-engineer.md`](agents\u002Fdata-ai\u002Ffeature-engineer.md) | Feature stores, pipelines, encoding strategies |\n| AutoResearch Agent | [`autoresearch-agent.md`](agents\u002Fdata-ai\u002Fautoresearch-agent.md) | ML experiment automation via tree search, code optimization |\n\n### Developer Experience (15 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| CLI Developer | [`cli-developer.md`](agents\u002Fdeveloper-experience\u002Fcli-developer.md) | CLI tools with Commander, yargs, clap |\n| DX Optimizer | [`dx-optimizer.md`](agents\u002Fdeveloper-experience\u002Fdx-optimizer.md) | Developer experience, tooling, ergonomics |\n| Documentation Engineer | [`documentation-engineer.md`](agents\u002Fdeveloper-experience\u002Fdocumentation-engineer.md) | Technical writing, API docs, guides |\n| Build Engineer | [`build-engineer.md`](agents\u002Fdeveloper-experience\u002Fbuild-engineer.md) | Build systems, bundlers, compilation |\n| Dependency Manager | [`dependency-manager.md`](agents\u002Fdeveloper-experience\u002Fdependency-manager.md) | Dependency audit, updates, lockfiles |\n| Refactoring Specialist | [`refactoring-specialist.md`](agents\u002Fdeveloper-experience\u002Frefactoring-specialist.md) | Code restructuring, dead code removal |\n| Legacy Modernizer | [`legacy-modernizer.md`](agents\u002Fdeveloper-experience\u002Flegacy-modernizer.md) | Legacy codebase migration strategies |\n| MCP Developer | [`mcp-developer.md`](agents\u002Fdeveloper-experience\u002Fmcp-developer.md) | MCP server and tool development |\n| Tooling Engineer | [`tooling-engineer.md`](agents\u002Fdeveloper-experience\u002Ftooling-engineer.md) | ESLint, Prettier, custom tooling |\n| Git Workflow Manager | [`git-workflow-manager.md`](agents\u002Fdeveloper-experience\u002Fgit-workflow-manager.md) | Branching strategies, CI, CODEOWNERS |\n| API Documentation | [`api-documentation.md`](agents\u002Fdeveloper-experience\u002Fapi-documentation.md) | OpenAPI\u002FSwagger, Redoc, interactive examples |\n| Monorepo Tooling | [`monorepo-tooling.md`](agents\u002Fdeveloper-experience\u002Fmonorepo-tooling.md) | Changesets, workspace deps, version management |\n| VS Code Extension | [`vscode-extension.md`](agents\u002Fdeveloper-experience\u002Fvscode-extension.md) | LSP integration, custom editors, webview panels |\n| Testing Infrastructure | [`testing-infrastructure.md`](agents\u002Fdeveloper-experience\u002Ftesting-infrastructure.md) | Test runners, CI splitting, flaky test management |\n| Developer Portal | [`developer-portal.md`](agents\u002Fdeveloper-experience\u002Fdeveloper-portal.md) | Backstage, service catalogs, self-service infra |\n\n### Specialized Domains (15 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| Blockchain Developer | [`blockchain-developer.md`](agents\u002Fspecialized-domains\u002Fblockchain-developer.md) | Smart contracts, Solidity, Web3 |\n| Game Developer | [`game-developer.md`](agents\u002Fspecialized-domains\u002Fgame-developer.md) | Game logic, ECS, state machines |\n| Embedded Systems | [`embedded-systems.md`](agents\u002Fspecialized-domains\u002Fembedded-systems.md) | Firmware, RTOS, hardware interfaces |\n| Fintech Engineer | [`fintech-engineer.md`](agents\u002Fspecialized-domains\u002Ffintech-engineer.md) | Financial systems, compliance, precision |\n| IoT Engineer | [`iot-engineer.md`](agents\u002Fspecialized-domains\u002Fiot-engineer.md) | MQTT, edge computing, digital twins |\n| Payment Integration | [`payment-integration.md`](agents\u002Fspecialized-domains\u002Fpayment-integration.md) | Stripe, PCI DSS, 3D Secure |\n| SEO Specialist | [`seo-specialist.md`](agents\u002Fspecialized-domains\u002Fseo-specialist.md) | Structured data, Core Web Vitals |\n| E-Commerce Engineer | [`e-commerce-engineer.md`](agents\u002Fspecialized-domains\u002Fe-commerce-engineer.md) | Cart, inventory, order management |\n| Healthcare Engineer | [`healthcare-engineer.md`](agents\u002Fspecialized-domains\u002Fhealthcare-engineer.md) | HIPAA, HL7 FHIR, medical data pipelines |\n| Real Estate Tech | [`real-estate-tech.md`](agents\u002Fspecialized-domains\u002Freal-estate-tech.md) | MLS integration, geospatial search, valuations |\n| Education Tech | [`education-tech.md`](agents\u002Fspecialized-domains\u002Feducation-tech.md) | LMS, SCORM\u002FxAPI, adaptive learning, assessments |\n| Media Streaming | [`media-streaming.md`](agents\u002Fspecialized-domains\u002Fmedia-streaming.md) | HLS\u002FDASH, transcoding, CDN, adaptive bitrate |\n| Geospatial Engineer | [`geospatial-engineer.md`](agents\u002Fspecialized-domains\u002Fgeospatial-engineer.md) | PostGIS, spatial queries, mapping APIs, tiles |\n| Robotics Engineer | [`robotics-engineer.md`](agents\u002Fspecialized-domains\u002Frobotics-engineer.md) | ROS2, sensor fusion, motion planning, SLAM |\n| Voice Assistant | [`voice-assistant.md`](agents\u002Fspecialized-domains\u002Fvoice-assistant.md) | STT, TTS, dialog management, Alexa\u002FGoogle |\n\n### Business & Product (12 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| Product Manager | [`product-manager.md`](agents\u002Fbusiness-product\u002Fproduct-manager.md) | PRDs, user stories, RICE prioritization |\n| Technical Writer | [`technical-writer.md`](agents\u002Fbusiness-product\u002Ftechnical-writer.md) | Documentation, style guides |\n| UX Researcher | [`ux-researcher.md`](agents\u002Fbusiness-product\u002Fux-researcher.md) | Usability testing, survey design |\n| Project Manager | [`project-manager.md`](agents\u002Fbusiness-product\u002Fproject-manager.md) | Sprint planning, Agile, task tracking |\n| Scrum Master | [`scrum-master.md`](agents\u002Fbusiness-product\u002Fscrum-master.md) | Ceremonies, velocity, retrospectives |\n| Business Analyst | [`business-analyst.md`](agents\u002Fbusiness-product\u002Fbusiness-analyst.md) | Requirements analysis, process mapping |\n| Content Strategist | [`content-strategist.md`](agents\u002Fbusiness-product\u002Fcontent-strategist.md) | SEO content, editorial calendars, topic clustering |\n| Growth Engineer | [`growth-engineer.md`](agents\u002Fbusiness-product\u002Fgrowth-engineer.md) | A\u002FB testing, analytics, funnel optimization |\n| Customer Success | [`customer-success.md`](agents\u002Fbusiness-product\u002Fcustomer-success.md) | Ticket triage, knowledge base, health scoring |\n| Sales Engineer | [`sales-engineer.md`](agents\u002Fbusiness-product\u002Fsales-engineer.md) | Technical demos, POCs, integration guides |\n| Legal Advisor | [`legal-advisor.md`](agents\u002Fbusiness-product\u002Flegal-advisor.md) | ToS, privacy policies, software licenses |\n| Marketing Analyst | [`marketing-analyst.md`](agents\u002Fbusiness-product\u002Fmarketing-analyst.md) | Campaign analysis, attribution, ROI tracking |\n\n### Orchestration (8 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| Task Coordinator | [`task-coordinator.md`](agents\u002Forchestration\u002Ftask-coordinator.md) | Routes work between agents, manages handoffs |\n| Context Manager | [`context-manager.md`](agents\u002Forchestration\u002Fcontext-manager.md) | Context compression, session summaries |\n| Workflow Director | [`workflow-director.md`](agents\u002Forchestration\u002Fworkflow-director.md) | Multi-agent pipeline orchestration |\n| Agent Installer | [`agent-installer.md`](agents\u002Forchestration\u002Fagent-installer.md) | Install and configure agent collections |\n| Knowledge Synthesizer | [`knowledge-synthesizer.md`](agents\u002Forchestration\u002Fknowledge-synthesizer.md) | Compress info, build knowledge graphs |\n| Performance Monitor | [`performance-monitor.md`](agents\u002Forchestration\u002Fperformance-monitor.md) | Track token usage, measure response quality |\n| Error Coordinator | [`error-coordinator.md`](agents\u002Forchestration\u002Ferror-coordinator.md) | Handle errors across multi-agent workflows |\n| Multi-Agent Coordinator | [`multi-agent-coordinator.md`](agents\u002Forchestration\u002Fmulti-agent-coordinator.md) | Parallel agent execution, merge outputs |\n\n### Research & Analysis (11 agents)\n\n| Agent | File | Purpose |\n|-------|------|---------|\n| Research Analyst | [`research-analyst.md`](agents\u002Fresearch-analysis\u002Fresearch-analyst.md) | Technical research, evidence synthesis |\n| Competitive Analyst | [`competitive-analyst.md`](agents\u002Fresearch-analysis\u002Fcompetitive-analyst.md) | Market positioning, feature comparison |\n| Trend Analyst | [`trend-analyst.md`](agents\u002Fresearch-analysis\u002Ftrend-analyst.md) | Technology trend forecasting |\n| Data Researcher | [`data-researcher.md`](agents\u002Fresearch-analysis\u002Fdata-researcher.md) | Data analysis, pattern recognition |\n| Search Specialist | [`search-specialist.md`](agents\u002Fresearch-analysis\u002Fsearch-specialist.md) | Information retrieval, source evaluation |\n| Patent Analyst | [`patent-analyst.md`](agents\u002Fresearch-analysis\u002Fpatent-analyst.md) | Patent searches, prior art, IP landscape |\n| Academic Researcher | [`academic-researcher.md`](agents\u002Fresearch-analysis\u002Facademic-researcher.md) | Literature reviews, citation analysis, methodology |\n| Market Researcher | [`market-researcher.md`](agents\u002Fresearch-analysis\u002Fmarket-researcher.md) | Market sizing, TAM\u002FSAM\u002FSOM, competitive intel |\n| Security Researcher | [`security-researcher.md`](agents\u002Fresearch-analysis\u002Fsecurity-researcher.md) | CVE analysis, threat modeling, attack surface |\n| Benchmarking Specialist | [`benchmarking-specialist.md`](agents\u002Fresearch-analysis\u002Fbenchmarking-specialist.md) | Performance benchmarks, comparative evals |\n| Technology Scout | [`technology-scout.md`](agents\u002Fresearch-analysis\u002Ftechnology-scout.md) | Emerging tech evaluation, build-vs-buy analysis |\n| agntk | [npx agntk](https:\u002F\u002Fgithub.com\u002FPhoenixrr2113\u002Fagntk) | Zero-config AI agent CLI with 20+ tools, persistent memory, Ollama auto-detection, and free tier |\n\n### Using Agents\n\nReference an agent in your `CLAUDE.md`:\n\n```markdown\n## Agents\n- Use `agents\u002Fcore-development\u002Ffullstack-engineer.md` for feature development\n- Use `agents\u002Fquality-assurance\u002Fcode-reviewer.md` for PR reviews\n- Use `agents\u002Fdata-ai\u002Fprompt-engineer.md` for prompt optimization\n```\n\n---\n\n## Skills\n\nThirty-five curated skill modules included in this repo, with access to **400,000+ additional skills** via the [SkillKit marketplace](https:\u002F\u002Fagenstskills.com). Each included skill teaches Claude Code domain-specific patterns with code examples, anti-patterns, and checklists.\n\n| Skill | Directory | What It Teaches |\n|-------|-----------|-----------------|\n| TDD Mastery | `skills\u002Ftdd-mastery\u002F` | Red-green-refactor, test-first design, coverage targets |\n| API Design Patterns | `skills\u002Fapi-design-patterns\u002F` | RESTful conventions, versioning, pagination, error responses |\n| Database Optimization | `skills\u002Fdatabase-optimization\u002F` | Query planning, indexing, N+1 prevention, connection pooling |\n| Frontend Excellence | `skills\u002Ffrontend-excellence\u002F` | Component architecture, state management, performance budgets |\n| Security Hardening | `skills\u002Fsecurity-hardening\u002F` | Input validation, auth patterns, secrets management, CSP |\n| [Prism Scanner](https:\u002F\u002Fgithub.com\u002Faidongise-cell\u002Fprism-scanner) | `pip install prism-scanner` | Agent skill\u002Fplugin\u002FMCP security scanner — 39+ rules, AST taint tracking, A-F grading |\n| DevOps Automation | `skills\u002Fdevops-automation\u002F` | Infrastructure as code, GitOps, monitoring, incident response |\n| Continuous Learning | `skills\u002Fcontinuous-learning\u002F` | Session summaries, learning logs, pattern extraction |\n| React Patterns | `skills\u002Freact-patterns\u002F` | Hooks, server components, suspense, error boundaries |\n| Python Best Practices | `skills\u002Fpython-best-practices\u002F` | Type hints, dataclasses, async\u002Fawait, packaging |\n| Go Idioms | `skills\u002Fgolang-idioms\u002F` | Error handling, interfaces, concurrency, project layout |\n| Django Patterns | `skills\u002Fdjango-patterns\u002F` | DRF, ORM optimization, signals, middleware |\n| Spring Boot Patterns | `skills\u002Fspringboot-patterns\u002F` | JPA, REST controllers, layered architecture |\n| Next.js Mastery | `skills\u002Fnextjs-mastery\u002F` | App Router, RSC, ISR, server actions, middleware |\n| GraphQL Design | `skills\u002Fgraphql-design\u002F` | Schema design, DataLoader, subscriptions, pagination |\n| Kubernetes Operations | `skills\u002Fkubernetes-operations\u002F` | Deployments, Helm charts, HPA, troubleshooting |\n| Docker Best Practices | `skills\u002Fdocker-best-practices\u002F` | Multi-stage builds, compose, image optimization |\n| AWS Cloud Patterns | `skills\u002Faws-cloud-patterns\u002F` | Lambda, DynamoDB, CDK, S3 event processing |\n| CI\u002FCD Pipelines | `skills\u002Fci-cd-pipelines\u002F` | GitHub Actions, GitLab CI, matrix builds |\n| Microservices Design | `skills\u002Fmicroservices-design\u002F` | Event-driven architecture, saga pattern, service mesh |\n| TypeScript Advanced | `skills\u002Ftypescript-advanced\u002F` | Generics, conditional types, mapped types, discriminated unions |\n| Rust Systems | `skills\u002Frust-systems\u002F` | Ownership, traits, async patterns, error handling |\n| Prompt Engineering | `skills\u002Fprompt-engineering\u002F` | Chain-of-thought, few-shot, structured outputs |\n| MCP Development | `skills\u002Fmcp-development\u002F` | MCP server tools, resources, transport setup |\n| PostgreSQL Optimization | `skills\u002Fpostgres-optimization\u002F` | EXPLAIN ANALYZE, indexes, partitioning, JSONB |\n| Redis Patterns | `skills\u002Fredis-patterns\u002F` | Caching, rate limiting, pub\u002Fsub, streams, Lua scripts |\n| Monitoring & Observability | `skills\u002Fmonitoring-observability\u002F` | OpenTelemetry, Prometheus, structured logging |\n| Authentication Patterns | `skills\u002Fauthentication-patterns\u002F` | JWT, OAuth2 PKCE, RBAC, session management |\n| WebSocket & Realtime | `skills\u002Fwebsocket-realtime\u002F` | Socket.io, SSE, reconnection, scaling |\n| Testing Strategies | `skills\u002Ftesting-strategies\u002F` | Contract testing, snapshot testing, property-based testing |\n| Git Advanced | `skills\u002Fgit-advanced\u002F` | Worktrees, bisect, interactive rebase, hooks |\n| Accessibility (WCAG) | `skills\u002Faccessibility-wcag\u002F` | ARIA patterns, keyboard navigation, color contrast |\n| Performance Optimization | `skills\u002Fperformance-optimization\u002F` | Code splitting, image optimization, Core Web Vitals |\n| Mobile Development | `skills\u002Fmobile-development\u002F` | React Native, Flutter, responsive layouts |\n| Data Engineering | `skills\u002Fdata-engineering\u002F` | ETL pipelines, Spark, star schema, data quality |\n| LLM Integration | `skills\u002Fllm-integration\u002F` | Streaming, function calling, RAG, cost optimization |\n\n### Community Skills\n\n| Skill | Install | What It Teaches |\n|-------|---------|------------------|\n| [Reepl - LinkedIn Content Creation](https:\u002F\u002Fgithub.com\u002Freepl-io\u002Fskills) | `npx skillkit@latest install reepl-io\u002Fskills` | 18 tools for LinkedIn content management: drafts, publishing, scheduling, voice profiles, contacts, collections, templates, and AI image generation |\n| [avoid-ai-writing](https:\u002F\u002Fgithub.com\u002Fconorbronsdon\u002Favoid-ai-writing) | `git clone https:\u002F\u002Fgithub.com\u002Fconorbronsdon\u002Favoid-ai-writing ~\u002F.claude\u002Fskills\u002Favoid-ai-writing` | AI writing pattern detection and rewriting (21 categories, 43 replacements) |\n| [CCM](https:\u002F\u002Fgithub.com\u002Fdr5hn\u002Fccm) | `npx skills add dr5hn\u002Fccm@ccm` | Multi-account management, session cleanup, health checks, environment snapshots, permissions audit for Claude Code |\n| [cc-inspect](https:\u002F\u002Fgithub.com\u002Fhowardpen9\u002Fcc-inspect) | `git clone https:\u002F\u002Fgithub.com\u002Fhowardpen9\u002Fcc-inspect ~\u002F.claude\u002Fskills\u002Fcc-inspect` | Inspect all installed Claude Code skills, plugins, MCP servers, commands, and hooks in a browser dashboard. Scope-aware (user\u002Fproject\u002Flocal), zero dependencies (pure bash + python3), generates self-contained HTML |\n| [Edison](https:\u002F\u002Fgithub.com\u002Fkilnside\u002Fedison) | `git clone https:\u002F\u002Fgithub.com\u002Fkilnside\u002Fedison ~\u002F.claude\u002Fskills\u002Fedison` | Design decision skill — the phase between brainstorming and building. Research-first progressive deepening with self-executing specs. Three modes: Check, Explore, Audit |\n| [claude-handoff](https:\u002F\u002Fgithub.com\u002FREMvisual\u002Fclaude-handoff) | `git clone https:\u002F\u002Fgithub.com\u002FREMvisual\u002Fclaude-handoff ~\u002F.claude\u002Fskills\u002Fhandoff` | Cross-session handoff system with chain tracking, context summaries, and structured work continuity between Claude Code sessions |\n| [ClawSec](https:\u002F\u002Fclawsec.cc) | Visit [clawsec.cc](https:\u002F\u002Fclawsec.cc) | Security audit platform for AI agent skills — automated 5-tier assessments, vulnerability detection, and configuration auditing for 33,000+ skills |\n| [ClawSearch](https:\u002F\u002Fclawsearch.cc) | `npm install -g clawsearch` | Security-first skill discovery engine with Trust Score ratings, knowledge graph, and pre-install guard (`clawsearch-guard`) to vet skills before installation |\n| [MUSE](https:\u002F\u002Fgithub.com\u002Fmyths-labs\u002Fmuse) | `git clone` + `.\u002Fscripts\u002Finstall.sh --tool claude` | Pure-Markdown memory OS with 48 skills, cross-conversation memory, 8 roles with permission isolation. Works with 6 AI coding tools. |\n| [PM Pilot](https:\u002F\u002Fgithub.com\u002Fmshadmanrahman\u002Fpm-pilot) | `git clone https:\u002F\u002Fgithub.com\u002Fmshadmanrahman\u002Fpm-pilot ~\u002F.claude\u002Fskills\u002Fpm-pilot` | Claude Code for PMs: 25 skills for meeting prep, PRDs, stakeholder intel, user story mapping, customer journey maps, and product discovery |\n| [Product Manager Skills](https:\u002F\u002Fgithub.com\u002FDigidai\u002Fproduct-manager-skills) | `clawhub install product-manager-skills` | Senior PM agent with 6 knowledge domains, 12 templates, and 30+ frameworks covering discovery, strategy, delivery, SaaS metrics, PM career coaching (IC to CPO), and AI product craft |\n| [OpenDivination](https:\u002F\u002Fgithub.com\u002Famenti-labs\u002Fopendivination) | `pipx install opendivination && npx skills add amenti-labs\u002Fopendivination --skill divination` | Tarot and I Ching skill with auditable entropy provenance, guided source setup across computer RNG, QRNG APIs, and local hardware, plus optional resonance mode |\n| [pumpclaw](https:\u002F\u002Fgithub.com\u002Fchainstacklabs\u002Fpumpclaw) | `git clone https:\u002F\u002Fgithub.com\u002Fchainstacklabs\u002Fpumpclaw ~\u002F.claude\u002Fskills\u002Fpumpclaw` | Agent skill for pump.fun token trading on Solana — buy, sell, launch tokens, wallet management, dry-run simulation, slippage checks, and PumpSwap AMM migrations via pumpfun-cli |\n| [x-twitter-scraper](https:\u002F\u002Fgithub.com\u002FXquik-dev\u002Fx-twitter-scraper) | `npx skills add Xquik-dev\u002Fx-twitter-scraper` | X API & Twitter scraper skill for AI coding agents -- tweet search, user lookup, follower extraction, engagement metrics, giveaway draws, trending topics, account monitoring, and 19 extraction tools |\n| [claude-code-marketing-skills](https:\u002F\u002Fgithub.com\u002Fcognyai\u002Fclaude-code-marketing-skills) | `git clone https:\u002F\u002Fgithub.com\u002Fcognyai\u002Fclaude-code-marketing-skills ~\u002F.claude\u002Fskills\u002Fclaude-code-marketing-skills` | AI-powered marketing skills — SEO Audit, Landing Page Review, Competitor Analysis, Ad Copy Writer, Lead Qualification. Free, no account required. Works with Claude Code, Cursor, Windsurf |\n| [claude-skills](https:\u002F\u002Fgithub.com\u002Falirezarezvani\u002Fclaude-skills) | `git clone` | 192 production-ready skills across 9 domains (engineering, marketing, product, compliance, C-level advisory) with 254 Python automation tools. 5,300+ stars |\n| [faf-skills](https:\u002F\u002Fgithub.com\u002FWolfe-Jam\u002Ffaf-skills) | `git clone` | 31 Claude Code skills for persistent project context (.faf, `application\u002Fvnd.faf+yaml`). Scoring, sync, testing, publishing, MCP server creation, architecture docs. Anthropic MCP #2759. |\n| [floom](https:\u002F\u002Fgithub.com\u002Ffloomhq\u002Ffloom) | `git clone https:\u002F\u002Fgithub.com\u002Ffloomhq\u002Ffloom.git ~\u002F.claude\u002Fskills\u002Ffloom-repo && ~\u002F.claude\u002Fskills\u002Ffloom-repo\u002Fscripts\u002Fsetup` | Deploy Python scripts as cloud automations. `\u002Ffloom` adapts code, tests in sandbox, deploys with auto-generated web UI, REST API, and MCP endpoint |\n| [n8n-skills](https:\u002F\u002Fgithub.com\u002Fczlonkowski\u002Fn8n-skills) | `git clone` | 7 complementary skills for building production-ready n8n workflows. Covers 525+ nodes, 2,653+ templates. 3,400+ stars |\n| [nv:context](https:\u002F\u002Fgithub.com\u002Fjohnnichev\u002Fnv-context) | `npx skills add johnnichev\u002Fnv-context -g -y` | State-of-the-art context engineering for AI agents. Auto-discovers tools, audits code for landmines, scores 6 leverage layers (repo structure, tool integration, memory, hooks, sessions, persona), generates production-ready CLAUDE.md\u002FAGENTS.md configs |\n| [claude-code-best-practice](https:\u002F\u002Fgithub.com\u002Fshanraisshan\u002Fclaude-code-best-practice) | `git clone` | Comprehensive reference implementation for Claude Code configuration -- skills, subagents, hooks, commands with practical examples. 17,400+ stars |\n| [ADHX](https:\u002F\u002Fgithub.com\u002Fitsmemeworks\u002Fadhx) | `\u002Fplugin marketplace add itsmemeworks\u002Fadhx` or `curl -sL https:\u002F\u002Fraw.githubusercontent.com\u002Fitsmemeworks\u002Fadhx\u002Fmain\u002Fskills\u002Fadhx\u002FSKILL.md -o ~\u002F.claude\u002Fskills\u002Fadhx\u002FSKILL.md` | Fetch any X\u002FTwitter post as clean LLM-friendly JSON — no scraping, works with tweets and full X Articles |\n| [SkillNav](https:\u002F\u002Fgithub.com\u002Fskillnav-dev\u002Fskillnav-skill) | `clawhub install HuiW86\u002Fskillnav` | Search 3,900+ MCP servers with install commands, get daily AI brief, query arXiv papers, and discover trending tools -- all in Chinese. Curated by skillnav.dev editorial team |\n| [moyu](https:\u002F\u002Fgithub.com\u002Fuucz\u002Fmoyu) | `claude skill install --url https:\u002F\u002Fgithub.com\u002Fuucz\u002Fmoyu --skill moyu` | Anti-over-engineering skill that teaches AI restraint. 5 variants (standard\u002Flite\u002Fstrict\u002Fen\u002Fja), 10 platforms. Benchmarked: 66% code reduction vs baseline |\n| [naming](https:\u002F\u002Fgithub.com\u002Fglacierphonk\u002Fnaming) | `git clone https:\u002F\u002Fgithub.com\u002Fglacierphonk\u002Fnaming ~\u002F.claude\u002Fskills\u002Fnaming` | Metaphor-driven naming for products, SaaS, brands, bots, and open source projects. Produces memorable, meaningful names |\n| [Obsidian Theme Designer](https:\u002F\u002Fgithub.com\u002FXiangyuSu611\u002Fobsidian-theme-designer) | `git clone https:\u002F\u002Fgithub.com\u002FXiangyuSu611\u002Fobsidian-theme-designer ~\u002F.claude\u002Fskills\u002Fobsidian-theme-designer` | Design Obsidian themes visually in the browser — style direction, color palette, font showcase, dual light\u002Fdark mode preview, and automated font installation. No CSS knowledge needed |\n| [Claudify](https:\u002F\u002Fclaudify.tech) | `npx create-claudify` | Complete operating system for Claude Code with 1,727 skills across 31 categories, 9 specialist agents with persistent memory, 21 slash commands, and automated quality checks |\n| [Axiom](https:\u002F\u002Fgithub.com\u002FGuipetris\u002Faxiom) | `npx skills add Guipetris\u002Faxiom` | Contract-enforced autonomous pipeline: 8 stages from idea to PR with EARS requirements, ATDD (tests before implementation), RALPLAN-DR deliberation with ADR, 3 independent parallel reviewers, classified self-correction (SYNTAX\u002FLOGIC\u002FARCH\u002FAMBIGUOUS), cross-run project memory, and stage rollback. Zero dependencies beyond git and gh. |\n| [AuraKit](https:\u002F\u002Fgithub.com\u002Fsmorky850612\u002FAurakit) | `npx @smorky85\u002Faurakit` | Fullstack development skill with 37 modes (BUILD\u002FFIX\u002FCLEAN\u002FDEPLOY\u002FREVIEW + 32 extended), 6-layer OWASP security enforcement, 30 runtime hooks, 10 language-specific reviewers, and an Instinct learning engine. Cross-platform: Claude Code, Codex CLI, Cursor, Windsurf, Manus. |\n| [StitchFlow](https:\u002F\u002Fgithub.com\u002Fyshishenya\u002Fstitchflow) | `git clone https:\u002F\u002Fgithub.com\u002Fyshishenya\u002Fstitchflow && cd stitchflow && bash install.sh --target all` | Cross-agent Stitch UI design bundle for turning briefs and mockups into screens, variants, Tailwind-friendly HTML, and screenshots |\n\n| [Cost Optimizer](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-cost-optimizer) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-cost-optimizer` | Save 60-80% on AI token costs with smart model routing, context compression, heartbeat tuning, usage reports, and config generation |\n| [Overnight Worker](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-overnight-worker) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-overnight-worker` | Autonomous overnight work agent — assign tasks before sleep, get structured results by morning with smart task decomposition, web research, and push notifications |\n| [Smart PR Review](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-smart-pr-review) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-smart-pr-review` | Opinionated AI code reviewer with 6-layer deep review, Devil's Advocate mode, and standardized MUST FIX \u002F SHOULD FIX \u002F SUGGESTION output for TS\u002FJS, Python, Go, Rust |\n| [DevOps Agent](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-devops-agent) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-devops-agent` | One-click deploy, monitoring setup (Prometheus+Grafana), scheduled backups, and fault diagnosis with safety-first design including confirmation prompts, dry-run, and snapshot rollback |\n| [CN Content Matrix](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-cn-content-matrix) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-cn-content-matrix` | Chinese multi-platform content generator for Xiaohongshu, WeChat, Douyin, and Bilibili with true style transfer, compliance review, and sensitive word detection |\n| [D3.js Visualization](https:\u002F\u002Fgithub.com\u002Fchrisvoncsefalvay\u002Fclaude-d3js-skill) | `git clone` | Generate interactive D3.js data visualizations from Claude Code |\n| [Google Drive – Memyard](https:\u002F\u002Fgithub.com\u002Fzagmoai\u002Fpublic-google-drive) | `git clone https:\u002F\u002Fgithub.com\u002Fzagmoai\u002Fpublic-google-drive ~\u002F.claude\u002Fskills\u002Fpublic-google-drive` | Create and edit Google Docs and Sheets without sign-in. Documents hosted on Memyard with shareable links. No API keys or OAuth required |\n| [iOS Simulator](https:\u002F\u002Fgithub.com\u002Fconorluddy\u002Fios-simulator-skill) | `git clone` | Interact with iOS Simulator for mobile testing and screenshot capture |\n| [Playwright Automation](https:\u002F\u002Fgithub.com\u002Flackeyjb\u002Fplaywright-skill) | `git clone` | Browser automation skill for end-to-end testing and web interaction |\n| [threat-hunting-with-sigma-rules](https:\u002F\u002Fgithub.com\u002Fjthack\u002Fthreat-hunting-with-sigma-rules-skill) | `git clone` | Sigma-rule-based threat detection and security log analysis |\n| [FFUF Web Fuzzing](https:\u002F\u002Fgithub.com\u002Fjthack\u002Fffuf_claude_skill) | `git clone` | Web vulnerability fuzzing and security testing via FFUF |\n| [Markdown to EPUB](https:\u002F\u002Fgithub.com\u002Fsmerchek\u002Fclaude-epub-skill) | `git clone` | Convert markdown files into EPUB ebooks |\n| [CSV Data Summarizer](https:\u002F\u002Fgithub.com\u002Fcoffeefuelbump\u002Fcsv-data-summarizer-claude-skill) | `git clone` | Generate insights and summaries from CSV data files |\n| [tapestry](https:\u002F\u002Fgithub.com\u002Fmichalparkola\u002Ftapestry-skills-for-claude-code) | `git clone` | Knowledge networks, iterative learning, article extraction, and YouTube transcript processing |\n| [j4rk0r\u002Fclaude-skills](https:\u002F\u002Fgithub.com\u002Fj4rk0r\u002Fclaude-skills) | `npx skills add j4rk0r\u002Fclaude-skills --yes --global` | 3 expert-grade skills: skill-guard (9-layer security auditor), skill-advisor (smart routing with gap analysis), skill-learner (persistent error correction). All A+ 120\u002F120 on skill-judge |\n| [SciAgent-Skills](https:\u002F\u002Fgithub.com\u002Fjaechang-hits\u002FSciAgent-Skills) | `git clone https:\u002F\u002Fgithub.com\u002Fjaechang-hits\u002FSciAgent-Skills` | 197 life science skills for Claude Code covering genomics, proteomics, drug discovery, and more. BixBench 92.0% accuracy. |\n| [manage-skills](https:\u002F\u002Fgithub.com\u002Fumutbozdag\u002Fagent-skills-manager\u002Ftree\u002Fmain\u002Fskills\u002Fmanage-skills) | `cp -r skills\u002Fmanage-skills ~\u002F.claude\u002Fskills\u002Fmanage-skills` | Discover, create, edit, toggle, copy, move, and delete agent skills across 11 tools (Cursor, Claude, Windsurf, Copilot, Codex, Cline, Aider, Continue, Roo Code, Augment) from the terminal |\n\n### Installing Skills\n\n**Browse and install via SkillKit** (recommended):\n\n```bash\nnpx skillkit@latest install claude-code-toolkit\u002Ftdd-mastery\n```\n\n### 15,000+ Skills via SkillKit Marketplace\n\nThis toolkit includes 35 curated skills. For access to **400,000+ additional skills** across every domain, use [SkillKit](https:\u002F\u002Fagenstskills.com):\n\n```bash\nnpx skillkit@latest                    # Launch interactive TUI\nnpx skillkit@latest search \"react\"     # Search 400,000+ skills\nnpx skillkit@latest recommend          # AI-powered skill recommendations\n```\n\nBrowse the full marketplace at [agenstskills.com](https:\u002F\u002Fagenstskills.com). SkillKit supports 32+ AI coding agents including Claude Code, Cursor, Codex, Gemini CLI, and more.\n\n---\n\n## Commands\n\nForty-two slash commands organized into eight categories. Drop these into your project's `.claude\u002Fcommands\u002F` directory.\n\n### Git (7 commands)\n\n| Command | File | Description |\n|---------|------|-------------|\n| `\u002Fcommit` | [`commit.md`](commands\u002Fgit\u002Fcommit.md) | Generate conventional commit from staged changes |\n| `\u002Fpr-create` | [`pr-create.md`](commands\u002Fgit\u002Fpr-create.md) | Create PR with summary, test plan, and labels |\n| `\u002Fchangelog` | [`changelog.md`](commands\u002Fgit\u002Fchangelog.md) | Generate changelog from commit history |\n| `\u002Frelease` | [`release.md`](commands\u002Fgit\u002Frelease.md) | Create tagged release with auto-generated notes |\n| `\u002Fworktree` | [`worktree.md`](commands\u002Fgit\u002Fworktree.md) | Set up git worktrees for parallel development |\n| `\u002Ffix-issue` | [`fix-issue.md`](commands\u002Fgit\u002Ffix-issue.md) | Fix a GitHub issue by number |\n| `\u002Fpr-review` | [`pr-review.md`](commands\u002Fgit\u002Fpr-review.md) | Review a pull request with structured feedback |\n\n### Testing (6 commands)\n\n| Command | File | Description |\n|---------|------|-------------|\n| `\u002Ftdd` | [`tdd.md`](commands\u002Ftesting\u002Ftdd.md) | Test-driven development workflow |\n| `\u002Ftest-coverage` | [`test-coverage.md`](commands\u002Ftesting\u002Ftest-coverage.md) | Analyze coverage and suggest missing tests |\n| `\u002Fe2e` | [`e2e.md`](commands\u002Ftesting\u002Fe2e.md) | Generate end-to-end test scenarios |\n| `\u002Fintegration-test` | [`integration-test.md`](commands\u002Ftesting\u002Fintegration-test.md) | Generate integration tests for API endpoints |\n| `\u002Fsnapshot-test` | [`snapshot-test.md`](commands\u002Ftesting\u002Fsnapshot-test.md) | Generate snapshot\u002Fgolden file tests |\n| `\u002Ftest-fix` | [`test-fix.md`](commands\u002Ftesting\u002Ftest-fix.md) | Diagnose and fix failing tests |\n\n### Architecture (6 commands)\n\n| Command | File | Description |\n|---------|------|-------------|\n| `\u002Fplan` | [`plan.md`](commands\u002Farchitecture\u002Fplan.md) | Create implementation plan with risk assessment |\n| `\u002Frefactor` | [`refactor.md`](commands\u002Farchitecture\u002Frefactor.md) | Structured code refactoring workflow |\n| `\u002Fmigrate` | [`migrate.md`](commands\u002Farchitecture\u002Fmigrate.md) | Framework or library migration |\n| `\u002Fadr` | [`adr.md`](commands\u002Farchitecture\u002Fadr.md) | Write Architecture Decision Record |\n| `\u002Fdiagram` | [`diagram.md`](commands\u002Farchitecture\u002Fdiagram.md) | Generate Mermaid diagrams from code |\n| `\u002Fdesign-review` | [`design-review.md`](commands\u002Farchitecture\u002Fdesign-review.md) | Conduct structured design review |\n\n### Documentation (5 commands)\n\n| Command | File | Description |\n|---------|------|-------------|\n| `\u002Fdoc-gen` | [`doc-gen.md`](commands\u002Fdocumentation\u002Fdoc-gen.md) | Generate documentation from code |\n| `\u002Fupdate-codemap` | [`update-codemap.md`](commands\u002Fdocumentation\u002Fupdate-codemap.md) | Update project code map |\n| `\u002Fapi-docs` | [`api-docs.md`](commands\u002Fdocumentation\u002Fapi-docs.md) | Generate API docs from route handlers |\n| `\u002Fonboard` | [`onboard.md`](commands\u002Fdocumentation\u002Fonboard.md) | Create onboarding guide for new devs |\n| `\u002Fmemory-bank` | [`memory-bank.md`](commands\u002Fdocumentation\u002Fmemory-bank.md) | Update CLAUDE.md memory bank |\n\n### Security (5 commands)\n\n| Command | File | Description |\n|---------|------|-------------|\n| `\u002Faudit` | [`audit.md`](commands\u002Fsecurity\u002Faudit.md) | Run security audit on code and dependencies |\n| `\u002Fhardening` | [`hardening.md`](commands\u002Fsecurity\u002Fhardening.md) | Apply security hardening measures |\n| `\u002Fsecrets-scan` | [`secrets-scan.md`](commands\u002Fsecurity\u002Fsecrets-scan.md) | Scan for leaked secrets and credentials |\n| `\u002Fcsp` | [`csp.md`](commands\u002Fsecurity\u002Fcsp.md) | Generate Content Security Policy headers |\n| `\u002Fdependency-audit` | [`dependency-audit.md`](commands\u002Fsecurity\u002Fdependency-audit.md) | Audit dependencies for vulnerabilities |\n\n### Refactoring (5 commands)\n\n| Command | File | Description |\n|---------|------|-------------|\n| `\u002Fdead-code` | [`dead-code.md`](commands\u002Frefactoring\u002Fdead-code.md) | Find and remove dead code |\n| `\u002Fsimplify` | [`simplify.md`](commands\u002Frefactoring\u002Fsimplify.md) | Reduce complexity of current file |\n| `\u002Fextract` | [`extract.md`](commands\u002Frefactoring\u002Fextract.md) | Extract function, component, or module |\n| `\u002Frename` | [`rename.md`](commands\u002Frefactoring\u002Frename.md) | Rename symbol across the codebase |\n| `\u002Fcleanup` | [`cleanup.md`](commands\u002Frefactoring\u002Fcleanup.md) | Remove dead code and unused imports |\n\n### DevOps (5 commands)\n\n| Command | File | Description |\n|---------|------|-------------|\n| `\u002Fdockerfile` | [`dockerfile.md`](commands\u002Fdevops\u002Fdockerfile.md) | Generate optimized Dockerfile |\n| `\u002Fci-pipeline` | [`ci-pipeline.md`](commands\u002Fdevops\u002Fci-pipeline.md) | Generate CI\u002FCD pipeline config |\n| `\u002Fk8s-manifest` | [`k8s-manifest.md`](commands\u002Fdevops\u002Fk8s-manifest.md) | Generate Kubernetes manifests |\n| `\u002Fdeploy` | [`deploy.md`](commands\u002Fdevops\u002Fdeploy.md) | Deploy to configured environment |\n| `\u002Fmonitor` | [`monitor.md`](commands\u002Fdevops\u002Fmonitor.md) | Set up monitoring and alerting |\n\n### Workflow (3 commands)\n\n| Command | File | Description |\n|---------|------|-------------|\n| `\u002Fcheckpoint` | [`checkpoint.md`](commands\u002Fworkflow\u002Fcheckpoint.md) | Save session progress and context |\n| `\u002Fwrap-up` | [`wrap-up.md`](commands\u002Fworkflow\u002Fwrap-up.md) | End session with summary and learnings |\n| `\u002Forchestrate` | [`orchestrate.md`](commands\u002Fworkflow\u002Forchestrate.md) | Run multi-agent workflow pipeline |\n\n### Using Commands\n\nCopy to your project:\n\n```bash\ncp -r commands\u002F .claude\u002Fcommands\u002F\n```\n\nThen invoke in Claude Code:\n\n```\n\u002Fcommit\n\u002Ftdd src\u002Futils\u002Fparser.ts\n\u002Faudit\n\u002Forchestrate feature \"Add user authentication\"\n```\n\n---\n\n## Hooks\n\nTwenty hook scripts covering all eight Claude Code lifecycle events. Place `hooks.json` in your `.claude\u002F` directory.\n\n### Hook Scripts\n\n| Script | Trigger | Purpose |\n|--------|---------|---------|\n| `session-start.js` | SessionStart | Load project context, detect package manager |\n| `session-end.js` | SessionEnd | Save session state for next session |\n| `context-loader.js` | SessionStart | Load CLAUDE.md, git status, pending todos |\n| `learning-log.js` | SessionEnd | Extract and save session learnings |\n| `pre-compact.js` | PreCompact | Save important context before compaction |\n| [`smart-approve.py`](https:\u002F\u002Fgithub.com\u002Fliberzon\u002Fclaude-hooks) | PreToolUse (Bash) | Decompose compound bash commands (&&, \\|\\|, ;, \\|, $()) into sub-commands and check each against allow\u002Fdeny patterns |\n| `block-dev-server.js` | PreToolUse (Bash) | Block dev server commands outside tmux |\n| `pre-push-check.js` | PreToolUse (Bash) | Verify branch and remote before push |\n| `block-md-creation.js` | PreToolUse (Write) | Block unnecessary .md file creation |\n| `commit-guard.js` | PreToolUse (Bash) | Validate conventional commit messages |\n| `secret-scanner.js` | PreToolUse (Write\u002FEdit) | Block files containing secrets |\n| `post-edit-check.js` | PostToolUse (Write\u002FEdit) | Run linter after file edits |\n| `auto-test.js` | PostToolUse (Write\u002FEdit) | Run related tests after edits |\n| `type-check.js` | PostToolUse (Write\u002FEdit) | TypeScript type checking after edits |\n| `lint-fix.js` | PostToolUse (Write\u002FEdit) | Auto-fix lint issues |\n| `bundle-check.js` | PostToolUse (Bash) | Check bundle size after builds |\n| `suggest-compact.js` | PostToolUse (Bash) | Suggest compaction at edit intervals |\n| `stop-check.js` | Stop | Remind to run tests if code was modified |\n| `notification-log.js` | Notification | Log notifications for later review |\n| `prompt-check.js` | UserPromptSubmit | Detect vague prompts, suggest clarification |\n\n### Related SDKs\n\nIf you prefer a typed, npm-installable foundation for writing hooks rather than raw scripts:\n\n- [claude-code-hooks](https:\u002F\u002Fgithub.com\u002FPayshak\u002Fclaude-code-hooks) — TypeScript SDK with `defineHook()`, typed event payloads for all 5 hook events, response builders, and unit-testable `.handle()` method. Zero dependencies.\n\n### Installing Hooks\n\n```bash\ncp hooks\u002Fhooks.json .claude\u002Fhooks.json\ncp -r hooks\u002Fscripts\u002F .claude\u002Fhooks\u002Fscripts\u002F\n```\n\n### Quick Setup with cc-safe-setup\n\nInstall 8 production-tested safety hooks in one command — destructive command blocker, branch push protection, secret leak prevention, syntax validation, and more:\n\n```bash\nnpx cc-safe-setup\n```\n\nIncludes `--audit` (score your setup 0-100), `--scan` (detect tech stack, recommend hooks), and `--verify` (test each hook). See [cc-safe-setup](https:\u002F\u002Fgithub.com\u002Fyurukusa\u002Fcc-safe-setup) for details.\n\n---\n\n## Rules\n\nFifteen coding rules that enforce consistent patterns. Add to `.claude\u002Frules\u002F` or reference in `CLAUDE.md`.\n\n| Rule | File | What It Enforces |\n|------|------|-----------------|\n| Coding Style | [`coding-style.md`](rules\u002Fcoding-style.md) | Naming conventions, file organization, import ordering |\n| Git Workflow | [`git-workflow.md`](rules\u002Fgit-workflow.md) | Branching, commit format, PR process |\n| Testing | [`testing.md`](rules\u002Ftesting.md) | Test structure, coverage targets, mocking guidelines |\n| Security | [`security.md`](rules\u002Fsecurity.md) | Input validation, secrets, parameterized queries |\n| Performance | [`performance.md`](rules\u002Fperformance.md) | Lazy loading, caching, bundle optimization |\n| Documentation | [`documentation.md`](rules\u002Fdocumentation.md) | JSDoc for public APIs, inline comments policy |\n| Error Handling | [`error-handling.md`](rules\u002Ferror-handling.md) | Explicit handling, typed errors, no empty catch |\n| Agents | [`agents.md`](rules\u002Fagents.md) | Agent design patterns, handoff protocols |\n| API Design | [`api-design.md`](rules\u002Fapi-design.md) | REST conventions, status codes, versioning |\n| Accessibility | [`accessibility.md`](rules\u002Faccessibility.md) | WCAG 2.2, ARIA, semantic HTML |\n| Database | [`database.md`](rules\u002Fdatabase.md) | Query patterns, migrations, N+1 prevention |\n| Dependency Management | [`dependency-management.md`](rules\u002Fdependency-management.md) | Version pinning, audit, update policies |\n| Code Review | [`code-review.md`](rules\u002Fcode-review.md) | Review checklist, approval criteria |\n| Monitoring | [`monitoring.md`](rules\u002Fmonitoring.md) | Logging standards, metrics, alerting |\n| Naming | [`naming.md`](rules\u002Fnaming.md) | Naming conventions per language |\n\n---\n\n## Templates\n\n- **[claude-code-blueprint](https:\u002F\u002Fgithub.com\u002Ffaizkhairi\u002Fclaude-code-blueprint)** - Battle-tested reference architecture for Claude Code power users. Specialized agents with model hhtiering, natural-language skills, lifecycle hooks, path-scoped rules, starter presets, benchmarks, battle stories, and cross-tool mapping. Zero dependencies, MIT licensed.\n- **[The CLAUDE.md Bible](https:\u002F\u002Fechochime3.gumroad.com\u002Fl\u002Fclaudemd-bible)** - 25 stack-specific CLAUDE.md configs covering React, Next.js, FastAPI, Django, Svelte, Chrome Extensions, CLI tools, MCP servers, and more. Each config is 80-150 lines of tested rules for the specific patterns and pitfalls of each stack, plus a masterclass guide.\n- **[TemplateClaw](https:\u002F\u002Fgithub.com\u002Fjeromwolf\u002Ftemplateclaw)** - Developer template hub and project scaffolding tool packaged as a Claude Code Plugin. 32 production-ready templates across 6 categories (Landing Pages, Dashboards, UI Components, Dev Methodology, Project Setup, Refactoring). 7 slash commands including `\u002Ftemplateclaw`, `\u002Ftemplateclaw-landing`, `\u002Ftemplateclaw-dashboard`, `\u002Ftemplateclaw-ui`. Compatible with Claude Code, Codex, Gemini CLI, and Cursor. MIT licensed.\n\n\nSeven CLAUDE.md templates for different project types.\n\n| Template | File | Use Case |\n|----------|------|----------|\n| Minimal | [`minimal.md`](templates\u002Fclaude-md\u002Fminimal.md) | Small projects, scripts, quick prototypes |\n| Standard | [`standard.md`](templates\u002Fclaude-md\u002Fstandard.md) | Most projects -- covers preferences, rules, workflows |\n| Comprehensive | [`comprehensive.md`](templates\u002Fclaude-md\u002Fcomprehensive.md) | Large codebases with detailed conventions |\n| Monorepo | [`monorepo.md`](templates\u002Fclaude-md\u002Fmonorepo.md) | Turborepo\u002FNx monorepo with multiple packages |\n| Enterprise | [`enterprise.md`](templates\u002Fclaude-md\u002Fenterprise.md) | Large teams with compliance and SSO |\n| Python Project | [`python-project.md`](templates\u002Fclaude-md\u002Fpython-project.md) | FastAPI\u002FDjango Python projects |\n| Fullstack App | [`fullstack-app.md`](templates\u002Fclaude-md\u002Ffullstack-app.md) | Next.js + API fullstack applications |\n\n```bash\ncp templates\u002Fclaude-md\u002Fstandard.md CLAUDE.md\n```\n\n---\n\n## MCP Configs\n\nThirteen curated Model Context Protocol server configurations.\n\n| Config | File | Servers Included |\n|--------|------|-----------------|\n| Recommended | [`recommended.json`](mcp-configs\u002Frecommended.json) | 14 essential servers for general development |\n| Full Stack | [`fullstack.json`](mcp-configs\u002Ffullstack.json) | Filesystem, GitHub, Postgres, Redis, Puppeteer |\n| Kubernetes | [`kubernetes.json`](mcp-configs\u002Fkubernetes.json) | kubectl-mcp-server, Docker, GitHub |\n| Data Science | [`data-science.json`](mcp-configs\u002Fdata-science.json) | Jupyter, SQLite, PostgreSQL, Filesystem |\n| Frontend | [`frontend.json`](mcp-configs\u002Ffrontend.json) | Puppeteer, Figma, Storybook |\n| Crypto \u002F DeFi | [`crypto-defi.json`](mcp-configs\u002Fcrypto-defi.json) | defi-mcp, Filesystem, Fetch, Memory |\n| DevOps | [`devops.json`](mcp-configs\u002Fdevops.json) | AWS, Docker, GitHub, Terraform, Sentry |\n| Research | [`research.json`](mcp-configs\u002Fresearch.json) | BGPT scientific papers, Brave Search, Fetch, Memory, Filesystem |\n| Observability | [`observability.json`](mcp-configs\u002Fobservability.json) | Iris eval & observability for agent tracing, quality evaluation, and cost tracking |\n| Security | [`security.json`](mcp-configs\u002Fsecurity.json) | Ghidra reverse engineering, Snyk vulnerability scanning |\n| Design | [`design.json`](mcp-configs\u002Fdesign.json) | Figma design context, Blender 3D automation |\n| Workflow Automation | [`workflow-automation.json`](mcp-configs\u002Fworkflow-automation.json) | n8n workflow builder, Pipedream integration |\n| Mobile | [`mobile.json`](mcp-configs\u002Fmobile.json) | Android ADB automation, Xcode build tools |\n| Finance | [`finance.json`](mcp-configs\u002Ffinance.json) | Chart Library pattern intelligence, Fetch, Memory |\n\n---\n\n## Contexts\n\nFive context modes that configure Claude Code's behavior for different tasks.\n\n| Context | File | Focus |\n|---------|------|-------|\n| Development | [`dev.md`](contexts\u002Fdev.md) | Iterate fast, follow patterns, test alongside code |\n| Code Review | [`review.md`](contexts\u002Freview.md) | Check logic, security, edge cases |\n| Research | [`research.md`](contexts\u002Fresearch.md) | Evaluate tools, compare alternatives, document findings |\n| Debug | [`debug.md`](contexts\u002Fdebug.md) | Reproduce, hypothesize, fix root cause, regression test |\n| Deploy | [`deploy.md`](contexts\u002Fdeploy.md) | Pre-deploy checklist, staging-first, rollback criteria |\n\n---\n\n## Examples\n\nThree walkthrough examples demonstrating real toolkit usage.\n\n| Example | File | Description |\n|---------|------|-------------|\n| Session Workflow | [`session-workflow.md`](examples\u002Fsession-workflow.md) | End-to-end productive development session |\n| Multi-Agent Pipeline | [`multi-agent-pipeline.md`](examples\u002Fmulti-agent-pipeline.md) | Chaining agents for a Stripe billing feature |\n| Project Setup | [`project-setup.md`](examples\u002Fproject-setup.md) | Setting up a new project with the full toolkit |\n\n---\n\n## Setup\n\n```bash\nbash setup\u002Finstall.sh\n```\n\nThe interactive installer clones the repo, symlinks configs, and installs plugins.\n\n---\n\n## Project Structure\n\n```\nclaude-code-toolkit\u002F               850+ files\n  plugins\u002F                         150+ plugins (220 command files + external)\n  agents\u002F                          135 agents across 10 categories\n    core-development\u002F              13 agents\n    language-experts\u002F              25 agents\n    infrastructure\u002F                11 agents\n    quality-assurance\u002F             10 agents\n    data-ai\u002F                       15 agents\n    developer-experience\u002F          15 agents\n    specialized-domains\u002F           15 agents\n    business-product\u002F              12 agents\n    orchestration\u002F                 8 agents\n    research-analysis\u002F             11 agents\n  skills\u002F                          35 SKILL.md files\n  commands\u002F                        42 commands across 8 categories\n  hooks\u002F\n    hooks.json                     25 hook entries\n    scripts\u002F                       19 Node.js scripts\n  rules\u002F                           15 coding rules\n  templates\u002Fclaude-md\u002F             7 CLAUDE.md templates\n  mcp-configs\u002F                     8 server configurations\n  contexts\u002F                        5 context modes\n  examples\u002F                        3 walkthrough examples\n  setup\u002F                           Interactive installer\n```\n\n---\n\n## Companion Apps & GUIs\n\n| Name | Stars | Description |\n|------|-------|-------------|\n| [ctrl](https:\u002F\u002Fctrl.bulletproof.sh) | - | Pixel-art office that visualizes AI coding agents (Claude Code, Codex, Gemini) working in real time. Standalone web app + daemon, relay sharing for remote access, multi-agent support |\n| [Opcode](https:\u002F\u002Fgithub.com\u002Fwinfunc\u002Fopcode) | 21,000+ | Tauri 2 desktop GUI and toolkit for Claude Code -- create custom agents with visual editor, usage analytics, MCP integration |\n| [CloudCLI](https:\u002F\u002Fgithub.com\u002Fsiteboon\u002Fclaudecodeui) | 8,400+ | Free open-source web\u002Fmobile UI for Claude Code, Cursor CLI, and Codex. Responsive design, integrated shell, file explorer, git explorer |\n| [Companion](https:\u002F\u002Fgithub.com\u002FThe-Vibe-Company\u002Fcompanion) | 2,200+ | Web & mobile UI for Claude Code & Codex. Launch parallel sessions, stream responses, approve tools, session recovery |\n| [CCHub](https:\u002F\u002Fgithub.com\u002FMoresl\u002Fcchub) | - | Tauri 2 desktop app for managing Claude Code ecosystem: MCP marketplace, config profiles, skills, workflows, hooks, security audit, autopilot. Cross-platform (~20MB) |\n| [Ruflo](https:\u002F\u002Fgithub.com\u002Fruvnet\u002Fruflo) | 21,200+ | Agent orchestration platform -- deploy multi-agent swarms, coordinate autonomous workflows, distributed swarm intelligence, RAG integration |\n| [Vibe Kanban](https:\u002F\u002Fgithub.com\u002FBloopAI\u002Fvibe-kanban) | 23,200+ | Kanban-based orchestration for 10+ coding agents with isolated git worktrees per agent |\n| [Claude-Code-Workflow](https:\u002F\u002Fgithub.com\u002Fcatlog22\u002FClaude-Code-Workflow) | 1,400+ | JSON-driven multi-agent cadence-team framework with intelligent CLI orchestration (Gemini\u002FQwen\u002FCodex) |\n| [ccswarm](https:\u002F\u002Fgithub.com\u002Fnwiizo\u002Fccswarm) | 127 | Rust-based multi-agent orchestration with specialized agent pools, Git worktree isolation, 93% token reduction |\n| [parallel-code](https:\u002F\u002Fgithub.com\u002Fjohannesjo\u002Fparallel-code) | 370+ | Run Claude Code, Codex, and Gemini side by side -- each in its own git worktree |\n| [Bernstein](https:\u002F\u002Fgithub.com\u002Fchernistry\u002Fbernstein) | 5+ | Python multi-agent orchestrator — spawns Claude Code, Codex CLI, and Gemini CLI in parallel on isolated git worktrees, verifies with tests, auto-commits. Zero LLM tokens on coordination |\n| [Poirot](https:\u002F\u002Fgithub.com\u002FLeonardoCardoso\u002FPoirot) | 96 | macOS app for browsing Claude Code sessions, viewing diffs, and re-running commands. Reads local transcripts, runs offline |\n| [TokenEater](https:\u002F\u002Fgithub.com\u002FAThevon\u002FTokenEater) | 179 | Native macOS menu bar app for monitoring Claude AI usage limits and watching coding sessions live |\n| [Claw](https:\u002F\u002Fgithub.com\u002Fjamesrochabrun\u002FClaw) | 86 | Native macOS app wrapping Claude Code SDK in Swift. Plan Mode, MCP Integration, Custom System Prompts |\n| [The Claude Protocol](https:\u002F\u002Fgithub.com\u002FAvivK5498\u002FThe-Claude-Protocol) | 149 | Enforcement layer wrapping Claude Code with 13 hooks -- blocks unsafe operations, enforces worktree isolation |\n| [Notch So Good](https:\u002F\u002Fgithub.com\u002Fdeepshal99\u002Fnotch-so-good) | new | macOS notch-based session monitor with pixel-art companion, 13 animations, smart notifications, multi-session support |\n| [Anima](https:\u002F\u002Fgithub.com\u002Fbtangonan\u002Fanima) | new | Native macOS companion for Claude Code. Per-project ASCII familiars, nim token economy, cross-session watcher daemon. Tauri v2 + Rust, 4MB binary |\n| [crit](https:\u002F\u002Fgithub.com\u002Ftomasz-tomczyk\u002Fcrit) | 105 | Local browser UI for inline code review of any file or agent output; integrates with Claude Code via plan-hook to review and approve plans before execution, outputs structured .crit.json for agent consumption. |\n| [telegram-ai-bridge](https:\u002F\u002Fgithub.com\u002FAliceLJY\u002Ftelegram-ai-bridge) | new | Run N parallel Claude Code sessions from your phone via Telegram. War Room mode, per-bot personas, A2A multi-agent collaboration, tool approval from phone. Supports Claude Code, Codex, and Gemini backends |\n| [Untether](https:\u002F\u002Fgithub.com\u002Flittlebearapps\u002Funtether) | 12 | Telegram bridge for Claude Code (and 5 other AI agents). Send tasks by voice or text, stream progress, approve changes with inline buttons from your phone |\n| [OctoAlly](https:\u002F\u002Fgithub.com\u002Fai-genius-automations\u002Foctoally) | 18 | AI coding session orchestration dashboard -- multi-agent hive-mind via RuFlo, Whisper voice dictation, tmux session persistence, built-in git source control, Electron desktop app |\n| [tokburn](https:\u002F\u002Fgithub.com\u002Flsvishaal\u002Ftokburn) | 3 | Local analytics dashboard for Claude Code -- calculates API-equivalent costs, detects waste patterns (repeated reads, floundering, cost outliers), serves web UI + REST API. Zero network calls, zero accounts, zero telemetry. `uvx tokburn serve`. |\n| [skill-builder](https:\u002F\u002Fgithub.com\u002FScottpedia0\u002Fskill-builder) | new | Analyzes your activity (shell history, git, browser, Claude threads) and generates working skills via LLM. 15 built-in skills + unlimited AI-generated. React UI with MCP server builder |\n| [claude-workspace-snapshot](https:\u002F\u002Fgithub.com\u002FREMvisual\u002Fclaude-workspace-snapshot) | new | Snapshot and restore live Claude Code sessions as named, color-coded Windows Terminal tabs. Detects running sessions via process inspection and .jsonl file activity. Windows only |\n| [KANBAII](https:\u002F\u002Fgithub.com\u002Fmartinmsaavedra\u002Fkanbaii) | new | AI-native kanban board for Claude Code — plan visually, track progress, let AI execute. Sequential engine (Ralph), parallel multi-agent (Teams), cost tracking, real-time dashboard. `npx kanbaii start` |\n| [Claude Session Visualizer](https:\u002F\u002Fgithub.com\u002Fanaypaul\u002Fclaude-session-visualizer) | new | Live visualization dashboard for Claude Code sessions with execution trees, token cost tracking, thinking trace exploration, and error debugging |\n| [CCHub](https:\u002F\u002Fgithub.com\u002FMoresl\u002Fcchub) | New | Tauri 2.0 desktop app for managing the full Claude Code ecosystem -- MCP servers, skills, hooks, config profiles. Auto-scans Claude Code, Claude Desktop, Cursor configs |\n| [AionUI](https:\u002F\u002Fgithub.com\u002FiOfficeAI\u002FAionUi) | 20,500+ | Free local open-source 24\u002F7 Cowork app for Gemini CLI, Claude Code, Codex, and OpenCode |\n| [chops](https:\u002F\u002Fgithub.com\u002FShpigford\u002Fchops) | 1,000+ | macOS app to browse, edit, and manage skills across Claude Code, Cursor, Codex, Windsurf, and Amp |\n| [ClaudeCode Launchpad CLI](https:\u002F\u002Fgithub.com\u002Fnoambrand\u002Fkivun-terminal) | 9 | Windows & macOS 2-minute installer for Claude Code — auto-installs Node.js, Git & Claude Code; two-line live status bar (model, context %, usage limits with reset timers); desktop shortcut with folder picker; right-click \"Open with ClaudeCode Launchpad CLI\" on any Windows folder; optional light-blue theme; Claude flags support |\n| [clideck](https:\u002F\u002Fgithub.com\u002Frustykuntz\u002Fclideck) | 31 | WhatsApp-like dashboard for managing AI coding agents (Claude Code, Codex, Gemini CLI, OpenCode) in one browser window. Live status, session resume, autopilot routing between agents, mobile remote |\n| [WhereMyTokens](https:\u002F\u002Fgithub.com\u002Fjeongwookie\u002FWhereMyTokens) | new | Windows system tray app for monitoring Claude Code token usage in real time. Per-session token counts, costs, context window progress, rate limit bars (5h\u002F1w), activity heatmaps, and model breakdowns. Registers as a Claude Code statusLine plugin for live data without polling. Windows only |\n| [Onepilot](https:\u002F\u002Fonepilotapp.com) | - | iOS app for running Claude Code, Codex, and other coding agents on remote servers via SSH from your phone. Full terminal emulator with SwiftTerm, agent session management, mobile access to dev machines |\n| [docker-claude-code](https:\u002F\u002Fgithub.com\u002Fgw0\u002Fdocker-claude-code) | new | Run Claude Code in an isolated Docker container with multi-profile support, security hardening, best-practice defaults, a set of pre-installed plugin\u002Fskill bundles and remote dev support. Drop-in replacement for `claude` — a simple shell alias is all it takes. |\n| [amux](https:\u002F\u002Fgithub.com\u002Fmixpeek\u002Famux) | new | Open-source agent multiplexer for running dozens of parallel Claude Code sessions with web dashboard, self-healing watchdog, kanban board, agent-to-agent REST API, and mobile PWA. Single Python file |\n| [cc-token-status](https:\u002F\u002Fgithub.com\u002Fechowonderfulworld\u002Fcc-token-status) | new | macOS menu bar dashboard for Claude Code. Shows plan usage limits, costs, tokens, trends, model breakdown, and multi-machine sync via iCloud. Single Python file, auto-updates |\n\n---\n\n## Ecosystem\n\nNotable projects, directories, and resources across the Claude Code ecosystem.\n\n| Name | Stars | Description |\n|------|-------|-------------|\n| [claude-mem](https:\u002F\u002Fgithub.com\u002Fthedotmack\u002Fclaude-mem) | 35,900+ | Auto-captures everything Claude does, compresses with AI, injects context into future sessions. #1 trending GitHub Feb 2026 |\n| [wshobson\u002Fagents](https:\u002F\u002Fgithub.com\u002Fwshobson\u002Fagents) | 31,300+ | 112 specialized agents, 16 orchestrators, 146 skills, 79 tools in 72 focused plugins |\n| [oh-my-claudecode](https:\u002F\u002Fgithub.com\u002FYeachan-Heo\u002Foh-my-claudecode) | 9,900+ | Teams-first multi-agent orchestration with 19 specialized agents and 28 skills |\n| [ccusage](https:\u002F\u002Fgithub.com\u002Fryoppippi\u002Fccusage) | 11,500+ | CLI for analyzing Claude Code usage from local JSONL files. Offline mode, zero API calls needed |\n| [cc-statistics](https:\u002F\u002Fgithub.com\u002FandroidZzT\u002Fcc-statistics) | 270+ | Three-in-one Claude Code stats: CLI + Web + native macOS SwiftUI panel. Token costs, code changes by language, efficiency scoring, weekly reports. Supports Codex and Cursor too |\n| [ccpm](https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fccpm) | 7,600+ | Project management with GitHub Issues + Git worktrees for parallel agent execution |\n| [claude-code-system-prompts](https:\u002F\u002Fgithub.com\u002FPiebald-AI\u002Fclaude-code-system-prompts) | 5,900+ | Extracted system prompts from Claude Code, updated for each release |\n| [claude-context](https:\u002F\u002Fgithub.com\u002Fzilliztech\u002Fclaude-context) | 5,600+ | Semantic code search MCP by Zilliz -- hybrid BM25 + vector search, ~40% token reduction |\n| [claude-skills](https:\u002F\u002Fgithub.com\u002Falirezarezvani\u002Fclaude-skills) | 5,300+ | 192 skills across 9 domains with 254 Python automation tools |\n| [paco-framework](https:\u002F\u002Fgithub.com\u002FPenguinAlleyApps\u002Fpaco-framework) | 1+ | Markdown-first multi-agent OS for Claude Code. Coordinates 3-16 specialized AI agents (Engineering, QA, Growth, Finance) with file-based dispatch, institutional memory, CEO Gate approval, and 24\u002F7 scheduling. No Python required. MIT |\n| [peon-ping](https:\u002F\u002Fgithub.com\u002FPeonPing\u002Fpeon-ping) | 3,900+ | Warcraft III Peon voice notifications for Claude Code and other agents |\n| [n8n-skills](https:\u002F\u002Fgithub.com\u002Fczlonkowski\u002Fn8n-skills) | 3,400+ | Skills for building production-ready n8n workflows |\n| [claude-hooks-mastery](https:\u002F\u002Fgithub.com\u002Fdisler\u002Fclaude-code-hooks-mastery) | 3,300+ | Complete mastery guide for Claude Code hooks with production-ready scripts |\n| [claude-supermemory](https:\u002F\u002Fgithub.com\u002Fsupermemoryai\u002Fclaude-supermemory) | 2,300+ | Persistent memory across sessions using Supermemory |\n| [myclaude](https:\u002F\u002Fgithub.com\u002Fstellarlinkco\u002Fmyclaude) | 2,400+ | Multi-agent orchestration routing to Claude Code, Codex, Gemini, and OpenCode |\n| [claude-code-mcp](https:\u002F\u002Fgithub.com\u002Fsteipete\u002Fclaude-code-mcp) | 1,100+ | Run Claude Code as a one-shot MCP server -- an agent in your agent |\n| [ccmanager](https:\u002F\u002Fgithub.com\u002Fkbwo\u002Fccmanager) | 940+ | Session manager supporting 8 coding agents with smart auto-approval |\n| [cog](https:\u002F\u002Fgithub.com\u002Fmarciopuga\u002Fcog) | 240+ | Cognitive architecture for Claude Code -- persistent memory, self-reflection, and foresight via plain-text conventions. Zero dependencies, just CLAUDE.md + markdown files |\n| [Cortex](https:\u002F\u002Fgithub.com\u002FSKULLFIRE07\u002Fcortex-memory) | -- | Persistent AI memory for coding assistants. Auto-captures decisions, patterns, and context across sessions. VSCode extension + CLI + MCP server. Free |\n| [openclaw-self-healing](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fopenclaw-self-healing) | 32+ | 4-tier autonomous crash recovery for Claude Code and any service — 64% auto-resolved, LLM-agnostic (Claude\u002FGPT-4\u002FGemini\u002FOllama), Prometheus metrics |\n| [CCM](https:\u002F\u002Fgithub.com\u002Fdr5hn\u002Fccm) | 8+ | Power-user toolkit for Claude Code — multi-account management with project bindings, session cleanup, environment snapshots, permissions audit, health diagnostics. Single bash script, no runtime dependencies beyond bash 4.4+ and jq. |\n| [openclaw-memorybox](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fopenclaw-memorybox) | 8+ | Memory hygiene CLI for Claude Code — prevents context overflow crashes, 83% MEMORY.md size reduction, zero dependencies |\n| [openclaw-self-evolving](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fopenclaw-self-evolving) | 2+ | Weekly self-improvement pipeline — scans Claude Code logs, proposes CLAUDE.md\u002FAGENTS.md rule changes, zero API cost |\n| [caliber](https:\u002F\u002Fgithub.com\u002Fcaliber-ai-org\u002Fai-setup) | -- | CLI that fingerprints projects and generates AI agent configs (CLAUDE.md, skills, AGENTS.md). Scores quality, auto-refreshes, supports Claude Code + Cursor + Codex |\n| [faf-cli](https:\u002F\u002Fgithub.com\u002FWolfe-Jam\u002Ffaf-cli) | -- | The package.json for AI context. IANA-registered `.faf` format — init, score, bi-sync with `CLAUDE.md`, export to AGENTS.md, GEMINI.md, .cursorrules. Built with Bun. |\n| [claude-starter-kit](https:\u002F\u002Fgithub.com\u002Fawrshift\u002Fclaude-starter-kit) | new | Ready-to-use project structure with persistent memory, session continuity, hooks, and 3 bundled skills (Gemini, Brainstorm, Design) |\n| [claude-code-kickstart](https:\u002F\u002Fgithub.com\u002Fypollak2\u002Fclaude-code-kickstart) | New | Opinionated starter kit — one command to install curated MCP servers, hooks, agents, and profiles. Includes auto-detect, 12 agents, 10 profiles, and 20+ shell commands |\n| [claude-code-power-stack](https:\u002F\u002Fgithub.com\u002Fbluzername\u002Fclaude-code-power-stack) | new | Ghost memory, conversation search, session naming, and Manus-style planning in a single install with cheat sheet PDF |\n| [clooks](https:\u002F\u002Fgithub.com\u002Fmauribadnights\u002Fclooks) | new | Persistent hook daemon that replaces per-invocation spawning -- 112x faster hooks with batching, dependency resolution, metrics |\n| [AIRIS MCP Gateway](https:\u002F\u002Fgithub.com\u002Fagiletec-inc\u002Fairis-mcp-gateway) | new | Docker-based MCP multiplexer that aggregates 60+ tools behind 7 meta-tools, reducing context token usage by 97%. One command to start, auto-enables servers on demand |\n| [Claude Code AWS Gateway](https:\u002F\u002Fgithub.com\u002Fantkawam\u002Fclaude-code-aws-gateway) | new | Self-hosted API gateway that routes Claude Code through Amazon Bedrock with team API keys, budgets, rate limits, OIDC SSO, SCIM provisioning, and an admin portal for analytics |\n| [gemini-claude-bridge](https:\u002F\u002Fgithub.com\u002Fweijiafu14\u002Fgemini-claude-bridge) | new | Gemini-to-Claude protocol converter for using Gemini models as Claude Code backend. Fixes 3 LiteLLM bugs |\n| [spartan-ai-toolkit](https:\u002F\u002Fgithub.com\u002Fspartan-stratos\u002Fspartan-ai-toolkit) | new | Engineering discipline layer for Claude Code -- 67 slash commands, 20 coding rules, 27 skills, 9 agents, quality gates between every step. 8 stack profiles (Go, Python, Java, Kotlin, React, etc.), agent memory across sessions. Install: `npx @c0x12c\u002Fspartan-ai-toolkit@latest --local` |\n| [AgenTopology](https:\u002F\u002Fgithub.com\u002Fagentopology\u002Fagentopology) | New | Declarative language and CLI for multi-agent orchestration -- define agents, flows, gates, hooks, group chats in one `.at` file, scaffold to 7 platforms, interactive visualizer. Apache 2.0 |\n| [cognitropy-lab](https:\u002F\u002Fgithub.com\u002FDaxxSec\u002Fcognitropy-lab) | -- | Autonomous workspace factory powered by a date-seeded entropy engine (363 domains, 22 categories, 18.8M unique outcomes). Builds one new Agent Workspace Model-compliant workspace daily — from firmware RE to aquaponics to satellite comms — with full slash commands, workflows, and secrets scanning. Zero human intervention.\n| [brood-box](https:\u002F\u002Fgithub.com\u002Fstacklok\u002Fbrood-box) | 10+ | Run AI coding agents (Claude Code, Codex, OpenCode) inside hardware-isolated microVMs with snapshot isolation, egress control, and MCP authorization |\n| [claude-kit](https:\u002F\u002Fgithub.com\u002Fluiseiman\u002Fclaude-kit) | 1 | Configuration factory for Claude Code -- 13 composable stacks, 6 agents, audit scoring, practices pipeline. Bootstraps and maintains .claude\u002F across projects |\n| [Global Chat](https:\u002F\u002Fgithub.com\u002Fpumanitro\u002Fglobal-chat) | -- | Cross-protocol MCP server discovery -- search 18K+ servers across 6+ registries. Also supports A2A and agents.txt. Free agents.txt validator. npm: `@global-chat\u002Fmcp-server` |\n| [GAAI Framework](https:\u002F\u002Fgithub.com\u002FFr-e-d\u002FGAAI-framework) | 95+ | Drop-in governance layer (.gaai\u002F folder) -- backlog authorization, cross-session memory, decision tracking, QA gates, autonomous delivery daemon. Claude Code + Cursor + Codex CLI + Gemini CLI |\n| [Wiggum CLI](https:\u002F\u002Fgithub.com\u002Ffedericoneri\u002Fwiggum-cli) | -- | Open-source AI agent that scans codebases (80+ tech), generates specs through AI interviews, and runs autonomous Ralph loops via Claude Code or Codex. Agent mode ships GitHub issues end-to-end with priority-aware scheduling and auto-merge |\n| [expert-dispatch](https:\u002F\u002Fgithub.com\u002Fsimonsysun\u002Fexpert-dispatch) | new | Specialist Dispatch Pattern -- let cheap AI assistants delegate complex tasks to Claude Code CLI. Bash adapter with run\u002Fresume\u002Freview workflow, session continuity, per-project directories, and structured JSON output. Designed to be called by another AI system |\n| [Claude Command Center](https:\u002F\u002Fgithub.com\u002Ftuning-labs-oss\u002Fclaude-command-center) | -- | Production-tested workspace template with session hooks, structured memory (4 types), multi-domain routing, daily task tracking, and interactive \u002Fsetup onboarding. 6 domain templates, pre-commit safety, example skills |\n| [Code Insights](https:\u002F\u002Fgithub.com\u002Fmelagiri\u002Fcode-insights) | 8 | Local-first CLI and dashboard for analyzing AI coding sessions from Claude Code, Cursor, Codex CLI, Copilot CLI, and VS Code Copilot. SQLite-backed with terminal analytics, browser dashboard, and LLM-powered insights |\n| [claude-code-cheat-sheet](https:\u002F\u002Fcc.storyfox.cz\u002F) | new | Complete one-page printable reference — all shortcuts, commands, CLI flags, MCP config, memory, skills, agents. Auto-updated daily from official docs. [EN](https:\u002F\u002Fcc.storyfox.cz\u002F) · [ZH](https:\u002F\u002Fcc.storyfox.cz\u002Fzh\u002F) · [JA](https:\u002F\u002Fcc.storyfox.cz\u002Fjp\u002F) · [KO](https:\u002F\u002Fcc.storyfox.cz\u002Fkr\u002F) |\n| [agent-dotfiles](https:\u002F\u002Fgithub.com\u002Fsaqibameen\u002Fagent-dotfiles) | new | Write AI coding rules once, sync to every agent. Supports Command Code, Claude Code, Cursor, Copilot, Codex, OpenCode. |\n| [zclean](https:\u002F\u002Fgithub.com\u002Fwhynowlab\u002Fzclean) | 19+ | Kills orphaned processes left behind by Claude Code and Codex. Stops zombie node\u002Fpython\u002Fesbuild from eating your RAM |\n| [Hindsight](https:\u002F\u002Fgithub.com\u002Fvectorize-io\u002Fhindsight) | -- | State-of-the-art long-term memory for AI agents by Vectorize. Biomimetic retain\u002Frecall\u002Freflect with 4 parallel retrieval strategies. Self-hosted or cloud, MIT-licensed. [Claude Code integration](https:\u002F\u002Fhindsight.vectorize.io\u002Fintegrations) |\n| [healthcare-agents](https:\u002F\u002Fgithub.com\u002Fajhcs\u002Fhealthcare-agents) | New | 51 specialized healthcare administration agents with MHA-level expertise across 10 divisions -- revenue cycle, compliance, quality, clinical ops, payer relations, health IT, and more |\n| [cc-switch](https:\u002F\u002Fgithub.com\u002Ffarion1231\u002Fcc-switch) | 35,500+ | Cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, and Gemini CLI |\n| [claude-code-router](https:\u002F\u002Fgithub.com\u002Fmusistudio\u002Fclaude-code-router) | 30,700+ | Use Claude Code as coding infrastructure foundation with custom model routing and interaction |\n| [SuperClaude](https:\u002F\u002Fgithub.com\u002FSuperClaude-Org\u002FSuperClaude_Framework) | 22,000+ | Config framework with specialized commands, cognitive personas, and dev methodologies |\n| [claude-code-templates](https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates) | 23,800+ | CLI tool for configuring and monitoring Claude Code projects and workflows |\n| [planning-with-files](https:\u002F\u002Fgithub.com\u002FOthmanAdi\u002Fplanning-with-files) | 17,600+ | Manus-style persistent markdown planning skill for structured project management |\n| [claude-hud](https:\u002F\u002Fgithub.com\u002Fjarrodwatts\u002Fclaude-hud) | 15,200+ | Plugin showing context usage, active tools, running agents, todo progress in a HUD overlay |\n| [codachi](https:\u002F\u002Fgithub.com\u002Fvincent-k2026\u002Fcodachi) | New | Tamagotchi-style statusline pet that grows with context usage, shows cache hit rate and burn speed |\n| [awesome-claude-code-subagents](https:\u002F\u002Fgithub.com\u002FVoltAgent\u002Fawesome-claude-code-subagents) | 15,600+ | 100+ specialized Claude Code subagents organized by domain |\n| [serena](https:\u002F\u002Fgithub.com\u002Foraios\u002Fserena) | 22,200+ | Semantic retrieval and editing MCP server for coding agents -- code-aware search and navigation |\n| [Understand-Anything](https:\u002F\u002Fgithub.com\u002FLum1104\u002FUnderstand-Anything) | 7,000+ | Turns any codebase into an interactive knowledge graph for exploration |\n| [n8n-mcp](https:\u002F\u002Fgithub.com\u002Fczlonkowski\u002Fn8n-mcp) | 17,000+ | MCP for Claude Code to build and manage n8n automation workflows |\n| [ToolRouter](https:\u002F\u002Ftoolrouter.com) | -- | Give Claude Code superpowers -- 150+ tools on demand with one account. Competitor research, video production, web search, image generation, security scanning, and more. `claude mcp add toolrouter -- npx -y toolrouter-mcp` |\n| [Claudoscope](https:\u002F\u002Fgithub.com\u002Fcordwainersmith\u002Fclaudoscope) | 29 | Native macOS menu bar app for Claude Code session analytics. Token\u002Fcost tracking, full-text search, real-time secret detection, config health linting (44 rules). Reads local JSONL files, fully offline, zero telemetry. Swift\u002FSwiftUI, Homebrew install |\n| [LynxPrompt](https:\u002F\u002Fgithub.com\u002FGeiserX\u002FLynxPrompt) | 25+ | Open-source platform for managing AI coding agent configs (CLAUDE.md, AGENTS.md, .cursorrules, copilot-instructions.md). Web marketplace, CLI, VS Code extension, self-hostable via Docker\u002FHelm with federation support |\n\n---\n\n## XVARY Stock Research\n\n- [XVARY Stock Research](https:\u002F\u002Fgithub.com\u002Fxvary-research\u002Fclaude-code-stock-analysis-skill) — Claude Code skill for public SEC EDGAR + market data: `\u002Fanalyze`, `\u002Fscore`, `\u002Fcompare`. MIT.\n\n---\n\n## Resources\n\nTutorials, guides, and deep-dives on Claude Code.\n\n| Resource | Type | Description |\n|----------|------|-------------|\n| [Claude Code Official Docs](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) | Docs | Official Anthropic documentation for Claude Code |\n| [Context Engineering Guide](https:\u002F\u002Fwww.anthropic.com\u002Fresearch\u002Flong-running-Claude) | Guide | Anthropic's guide to long-running Claude Code sessions and scientific computing |\n| [Auto Mode Engineering](https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fclaude-code-auto-mode) | Blog | How Auto Mode's two-layer safety system works under the hood |\n| [Claude Code Guide](https:\u002F\u002Fclaudecodeguide.dev) | Guide | Beginner-friendly guide from first install to daily operating system, zero jargon ([GitHub](https:\u002F\u002Fgithub.com\u002Fmshadmanrahman\u002Fclaudecode-guide)) |\n| [Claude Code Hooks Mastery](https:\u002F\u002Fgithub.com\u002Fdisler\u002Fclaude-code-hooks-mastery) | Tutorial | Complete mastery guide with production-ready hook scripts |\n| [Skills vs MCP vs Plugins](https:\u002F\u002Fwww.morphllm.com\u002Fcontext-engineering) | Guide | When to use which extension mechanism in Claude Code |\n| [Context Engineering 101](https:\u002F\u002Fnewsletter.victordibia.com\u002Fp\u002Fcontext-engineering-101-how-agents) | Article | Three core strategies: compaction, isolation, agentic memory |\n| [Agent Teams Explained](https:\u002F\u002Fwww.turingcollege.com\u002Fblog\u002Fclaude-agent-teams-explained) | Guide | How multiple Claude Code instances coordinate as an AI engineering team |\n| [84 Best Practices](https:\u002F\u002Fdiscuss.huggingface.co\u002Ft\u002F10-essential-claude-code-best-practices-you-need-to-know\u002F174731) | Collection | Community-compiled Claude Code best practices from Hugging Face |\n| [Computer Use from CLI](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fcomputer-use) | Docs | Let Claude control your desktop — open apps, click, type, screenshot. macOS, Pro\u002FMax, v2.1.85+. Per-app approval, tiered access, Esc abort |\n| [Dispatch & Remote Control](https:\u002F\u002Fclaude.com\u002Fblog\u002Fdispatch-and-computer-use) | Blog | Run Claude Code from phone\u002Fweb, schedule recurring tasks, remote session control |\n| [Claude Code Cheat Sheet](https:\u002F\u002Fcc.storyfox.cz\u002F) | Reference | One-page printable reference, auto-updated daily, 4 languages |\n| [Multi-Agent Orchestra](https:\u002F\u002Faddyosmani.com\u002Fblog\u002Fcode-agent-orchestra\u002F) | Blog | Addy Osmani's survey of multi-agent coding patterns |\n\n## Related Awesome Lists\n\n| List | Stars | Focus |\n|------|-------|-------|\n| [awesome-mcp-servers](https:\u002F\u002Fgithub.com\u002Fappcypher\u002Fawesome-mcp-servers) | 5,300+ | Curated MCP servers across 30 categories |\n| [awesome-ai-agents](https:\u002F\u002Fgithub.com\u002Fe2b-dev\u002Fawesome-ai-agents) | 27,000+ | Open-source and closed-source AI agents |\n| [awesome-llm-apps](https:\u002F\u002Fgithub.com\u002FShubhamsaboo\u002Fawesome-llm-apps) | 104,000+ | LLM-powered applications and multi-agent systems |\n| [awesome-claude-code (hesreallyhim)](https:\u002F\u002Fgithub.com\u002Fhesreallyhim\u002Fawesome-claude-code) | 34,000+ | Skills, hooks, slash-commands, and orchestrators |\n| [antigravity-awesome-skills](https:\u002F\u002Fgithub.com\u002Fsickn33\u002Fantigravity-awesome-skills) | 28,700+ | 1,326+ installable agentic skills with CLI |\n| [awesome-claude-code-subagents](https:\u002F\u002Fgithub.com\u002FVoltAgent\u002Fawesome-claude-code-subagents) | 15,600+ | 100+ specialized Claude Code subagents |\n| [claude-code-best-practice](https:\u002F\u002Fgithub.com\u002Fshanraisshan\u002Fclaude-code-best-practice) | 25,000+ | Comprehensive reference implementation for Claude Code config |\n| [awesome-claude-skills (Composio)](https:\u002F\u002Fgithub.com\u002FComposioHQ\u002Fawesome-claude-skills) | 49,300+ | 30 curated skills + 832 SaaS automation templates via Composio. Strong on document processing, creative, and business skills |\n\n## Contributing\n\nContributions are welcome. See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.\n\n## License\n\n[Apache-2.0](LICENSE)\n| [AutoSearch](https:\u002F\u002Fgithub.com\u002F0xmariowu\u002FAutosearch) | new | Self-evolving deep research plugin for Claude Code -- 32 dedicated search channels (arXiv, GitHub, Reddit, HN, zhihu, bilibili, CSDN + 25 more), LLM-scored evaluation, cited reports in Markdown\u002FHTML\u002FSlides, cross-session learning. Zero API keys required |\n","# Claude 代码工具包\n\n**Claude 代码最全面的工具包——包含135个智能体、35个精选技能（通过[SkillKit](https:\u002F\u002Fagenstskills.com)可扩展至40多万个）、42条命令、176多个插件、20个钩子、15条规则、7个模板、14个MCP配置、26个配套应用、52个生态项目等丰富内容。**\n\n[![Awesome](https:\u002F\u002Fcdn.rawgit.com\u002Fsindresorhus\u002Fawesome\u002Fd7305f38d29fed78fa85652e3a63e154dd8e8829\u002Fmedia\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fsindresorhus\u002Fawesome)\n[![许可证：Apache-2.0](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-Apache%202.0-blue.svg)](LICENSE)\n[![欢迎提交PR](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-welcome-brightgreen.svg)](CONTRIBUTING.md)\n[![最后更新](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLast%20Updated-Mar%202026-orange.svg)](#)\n[![文件数](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FFiles-850+-blueviolet.svg)](#project-structure)\n\n\u003Ca href=\"https:\u002F\u002Fstar-history.com\u002F#rohitg00\u002Fawesome-claude-code-toolkit&Date\">\n  \u003Cpicture>\n    \u003Csource media=\"(prefers-color-scheme: dark)\" srcset=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frohitg00_awesome-claude-code-toolkit_readme_d92a6d46027b.png&theme=dark\" \u002F>\n    \u003Csource media=\"(prefers-color-scheme: light)\" srcset=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frohitg00_awesome-claude-code-toolkit_readme_d92a6d46027b.png\" \u002F>\n    \u003Cimg alt=\"Star历史图表\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frohitg00_awesome-claude-code-toolkit_readme_d92a6d46027b.png\" width=\"600\" \u002F>\n  \u003C\u002Fpicture>\n\u003C\u002Fa>\n\n---\n\n## 快速安装\n\n**插件市场**（推荐）：\n\n```bash\n\u002Fplugin marketplace add rohitg00\u002Fawesome-claude-code-toolkit\n```\n\n**手动克隆：**\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit.git ~\u002F.claude\u002Fplugins\u002Fclaude-code-toolkit\n```\n\n**一行命令：**\n\n```bash\ncurl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fmain\u002Fsetup\u002Finstall.sh | bash\n```\n\n---\n\n## 目录\n\n- [插件](#plugins)（176+）\n- [智能体](#agents)（135）\n- [技能](#skills)（35个精选 + 28个社区贡献）\n- [命令](#commands)（42）\n- [钩子](#hooks)（20个脚本）\n- [规则](#rules)（15）\n- [模板](#templates)（7）\n- [MCP配置](#mcp-configs)（14）\n- [上下文](#contexts)（5）\n- [示例](#examples)（3）\n- [配套应用与GUI](#companion-apps--guis)\n- [生态系统](#ecosystem)\n- [设置](#setup)\n- [资源](#resources)\n- [相关Awesome列表](#related-awesome-lists)\n- [贡献](#contributing)\n\n---\n\n## 插件\n\n超过176个生产就绪的插件，为Claude Code扩展领域特定的能力。\n\n### 精选\n\n| 插件 | 星标 | 描述 |\n|--------|-------|-------------|\n| [pro-workflow](https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fpro-workflow) | 1,800+ | 经过实战检验的Claude Code工作流，来自资深用户。具备自我修正记忆、并行工作树、收尾仪式等功能，包含8种钩子类型、5个智能体以及80\u002F20的AI编码比例。安装方法：`\u002Fplugin marketplace add rohitg00\u002Fpro-workflow` |\n| [gstack](https:\u002F\u002Fgithub.com\u002Fgarrytan\u002Fgstack) | 68,200+ | Garry Tan的精确Claude Code配置：6款意见明确的工具，分别充当CEO、工程经理、发布经理和QA工程师的角色。 |\n\n### 所有插件\n\n| 插件 | 描述 |\n|--------|-------------|\n| [skills-janitor](https:\u002F\u002Fgithub.com\u002Fkhendzel\u002Fskills-janitor) | 审计、去重、检查、修复并跟踪你的 Claude Code 技能使用情况。9 个斜杠命令，零依赖 |\n| [aws-cost-saver](https:\u002F\u002Fgithub.com\u002Fprajapatimehul\u002Faws-cost-saver) | AWS 成本优化扫描器，包含 173 项自动化检查、基于机器学习的资源优化调整，以及“零幻觉”定价——实际效果：成本降低 60% |\n| [claude-agentic-coding-playbook](https:\u002F\u002Fgithub.com\u002Fjohn-wilmes\u002Fclaude-agentic-coding-playbook) | 基于证据的 LLM 辅助开发实践——钩子、技能、脚本，以及一本包含 58 篇引用的最佳实践指南。包括 19 个以上的守护\u002F生命周期钩子、调查工作流、舰队索引，以及用于自主任务队列的 claude-loop。 |\n| [claude-code-mcp](https:\u002F\u002Fgithub.com\u002Fsteipete\u002Fclaude-code-mcp) | 将 Claude Code 作为一次性 MCP 服务器运行——一个代理中的代理。权限自动绕过。作者：Peter Steinberger。1,100+ 颗星 |\n| [claude-code-system-prompts](https:\u002F\u002Fgithub.com\u002FPiebald-AI\u002Fclaude-code-system-prompts) | 从 Claude Code 中提取的系统提示——18 种内置工具描述、子代理提示、实用程序提示。每次发布都会更新。5,900+ 颗星 |\n| [claude-context](https:\u002F\u002Fgithub.com\u002Fzilliztech\u002Fclaude-context) | 由 Zilliz（Milvus 的创建者）开发的语义代码搜索 MCP 服务器。混合 BM25 和密集向量搜索。令牌消耗减少约 40%。5,600+ 颗星 |\n| [claude-cost-optimizer](https:\u002F\u002Fgithub.com\u002FSagargupta16\u002Fclaude-cost-optimizer) | 可安装的成本模式技能（节省 30–60%，严格模式下最高可达 70%）+ 10 份指南、10 个模板、预算挂钩。安装命令：`npx skills add Sagargupta16\u002Fclaude-cost-optimizer` |\n| [fractal](https:\u002F\u002Fgithub.com\u002Frmolines\u002Ffractal) | 针对 Claude Code 的递归式项目管理。将目标分解为可验证的谓词，优先处理风险最高的部分，并在学习过程中不断重新评估 |\n| [a11y-audit](plugins\u002Fa11y-audit\u002F) | 全面的无障碍审计，包含 WCAG 合规性检查 |\n| [accessibility-checker](plugins\u002Faccessibility-checker\u002F) | 扫描 Web 应用中的无障碍问题，并修复 ARIA 属性 |\n| [adr-writer](plugins\u002Fadr-writer\u002F) | 架构决策记录的编写与管理 |\n| [AgentLint](https:\u002F\u002Fgithub.com\u002F0xmariowu\u002FAgentLint) | 检查你的代码库是否适合 AI 代理使用。涵盖 5 个维度的 33 项基于证据的检查。Claude Code 插件。 |\n| [ai-prompt-lab](plugins\u002Fai-prompt-lab\u002F) | 改进和测试 AI 提示，以提升与 Claude Code 的交互效果 |\n| [analytics-reporter](plugins\u002Fanalytics-reporter\u002F) | 根据项目数据生成分析报告和仪表盘配置 |\n| [android-developer](plugins\u002Fandroid-developer\u002F) | 使用 Jetpack Compose 进行 Android 和 Kotlin 开发 |\n| [api-architect](plugins\u002Fapi-architect\u002F) | 使用 OpenAPI 规范生成 API 设计、文档和测试 |\n| [api-benchmarker](plugins\u002Fapi-benchmarker\u002F) | 对 API 端点进行基准测试和性能报告 |\n| [api-reference](plugins\u002Fapi-reference\u002F) | 从源代码生成 API 参考文档 |\n| [api-tester](plugins\u002Fapi-tester\u002F) | 测试 API 端点并对服务进行负载测试 |\n| [aws-helper](plugins\u002Faws-helper\u002F) | 自动化 AWS 服务配置和部署 |\n| [azure-helper](plugins\u002Fazure-helper\u002F) | 自动化 Azure 服务配置和部署 |\n| [backend-architect](plugins\u002Fbackend-architect\u002F) | 使用端点脚手架设计后端服务架构 |\n| [bug-detective](plugins\u002Fbug-detective\u002F) | 通过根本原因分析和执行跟踪系统地调试问题 |\n| [Bouncer](https:\u002F\u002Fgithub.com\u002Fbuildingopen\u002Fbouncer) | 独立的质量门，利用 Gemini 审计 Claude Code 的输出。包含自动停止钩子、快速审计技能以及具有完整工具访问权限的深度审计。一行命令即可安装。 |\n| [brooks-lint](https:\u002F\u002Fgithub.com\u002Fhyhmrright\u002Fbrooks-lint) | 基于六本经典工程书籍（Brooks、Fowler、Martin、McConnell、Hunt & Thomas、Evans）的 AI 代码审查。从六个退化风险维度诊断代码，并提供结构化的发现结果（症状 → 来源 → 结果 → 解决方案）。支持 PR 审查、架构审计、技术债务评估和测试质量审查。v0.6 新增了 Mermaid 依赖图。 |\n| [ccmanager](https:\u002F\u002Fgithub.com\u002Fkbwo\u002Fccmanager) | 编码代理会话管理器，支持 Claude Code、Gemini CLI、Codex、Cursor、Copilot、Cline、OpenCode 和 Kimi CLI。通过 Haiku 实现智能自动批准，并支持 DevContainer。940+ 颗星 |\n| [ccpm](https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fccpm) | 使用 GitHub Issues 和 Git 工作树进行并行代理执行的项目管理。包含 issue-analyze、epic-start 和 epic-merge 等命令。7,600+ 颗星 |\n| [ccusage](https:\u002F\u002Fgithub.com\u002Fryoppippi\u002Fccusage) | 用于从本地 JSONL 文件分析 Claude Code\u002FCodex 使用情况的命令行工具。提供每日、每月、会话及计费窗口报告。离线使用，无需调用任何 API。11,500+ 颗星 |\n| [building-multiagent-systems](https:\u002F\u002Fgithub.com\u002F2389-research\u002Fbuilding-multiagent-systems) | 多代理系统的架构模式：编排者-工作者、流水线、辩论以及 MapReduce 拓扑 |\n| [bundle-analyzer](plugins\u002Fbundle-analyzer\u002F) | 前端包大小分析和 Tree-Shaking 优化 |\n| [chief](https:\u002F\u002Fgithub.com\u002FMiniCodeMonkey\u002Fchief) | 将 Claude Code 包装在一个循环中的 CLI。定义 PRD，运行 chief，然后去做其他事情。每完成一项任务就会提交一次，接着从中断的地方继续。可通过 Homebrew 安装。380+ 颗星 |\n| [claudebase](https:\u002F\u002Fgithub.com\u002Frohithzr\u002Fclaudebase) | 备份、恢复和同步你的 Claude Code 配置（设置、技能、代理、钩子、规则、内存、MCP）到私有 GitHub 仓库。支持命名配置文件、秘密扫描、多机冲突检测以及自动备份。经过 158 次测试 |\n| [changelog-gen](plugins\u002Fchangelog-gen\u002F) | 通过解析常规提交记录生成变更日志 |\n| [changelog-writer](plugins\u002Fchangelog-writer\u002F) | 根据 Git 历史和 PR 编写详细的变更日志 |\n| [ci-debugger](plugins\u002Fci-debugger\u002F) | 调试 CI\u002FCD 流水线故障并修复配置 |\n| [claude-hooks-mastery](https:\u002F\u002Fgithub.com\u002Fdisler\u002Fclaude-code-hooks-mastery) | 关于 Claude Code 钩子的完整精通指南——UV 单文件 Python 脚本、子代理、元代理、团队验证机制以及 AI 生成的音频反馈。3,300+ 颗星 |\n| [claude-mem](https:\u002F\u002Fgithub.com\u002Fthedotmack\u002Fclaude-mem) | 自动捕获 Claude Code 的所有操作，利用 AI 进行压缩，并将相关上下文注入到未来的会话中。采用 SQLite 和全文搜索。35,900+ 颗星 |\n| [claude-notifications-go](https:\u002F\u002Fgithub.com\u002F777genius\u002Fclaude-notifications-go) | 跨平台智能通知——6 种类型、点击聚焦、上下文分析和 Webhook。单个 Go 二进制文件，无依赖。340+ 颗星 |\n| [claude-scaffold](https:\u002F\u002Fgithub.com\u002Fpyramidheadshark\u002Fclaude-scaffold) | 一个 npx CLI，可在一条命令内将 CLAUDE.md、钩子和 18 个领域技能部署到任何仓库中。技能会根据项目上下文通过钩子自动激活。可通过 `update --all` 实现跨仓库同步。安装命令：`npx claude-scaffold init` |\n| [claude-recap](https:\u002F\u002Fgithub.com\u002Fhatawong\u002Fclaude-recap) | 使用 Shell 钩子实现按主题的会话记忆——将每个对话主题存档为单独的 Markdown 摘要。两个钩子，分别基于 Bash 和 Node.js，完全本地化 |\n| [clirank-mcp-server](https:\u002F\u002Fgithub.com\u002Falexanderclapp\u002Fclirank-mcp-server)（[npm](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002Fclirank-mcp-server)）| 代理的 API 发现——根据 11 个信号对 210 多个 API 的 CLI 友好度进行评分。通过 MCP 工具或 REST (`curl \"https:\u002F\u002Fclirank.dev\u002Fapi\u002Fdiscover?q=send+emails\"`) 找到适合任何任务的正确 API。无需身份验证 |\n| [\u002Fcht](https:\u002F\u002Fgithub.com\u002Fkosoukhov\u002Fcht-cli) | 为 Claude Code 提供聊天持久化功能——13 个斜杠命令 + 钩子，可将对话保存为本地 Markdown 文件。通过钩子实现防崩溃的自动保存，按项目组织、搜索和管理。npm：`@kosoukhov\u002Fcht-cli` |\n| [claude-rank](https:\u002F\u002Fgithub.com\u002FHouseofmvps\u002Fclaude-rank) | SEO\u002FGEO\u002FAEO 审计插件——告诉你为什么 AI 不会引用你的网站，然后自动修复 robots.txt、sitemap.xml、llms.txt 和 JSON-LD。包含 10 个扫描器的 170 多条规则，无需配置 |\n| [claude-sounds](https:\u002F\u002Fgithub.com\u002FculminationAI\u002Fclaude-sounds) | 为 Claude Code 钩子提供音频反馈——10 种事件、21 种声音，随机轮换。适用于 macOS（`afplay`）。 |\n| [claude-code-hooks](https:\u002F\u002Fgithub.com\u002Fyurukusa\u002Fclaude-code-hooks) | 经过 160 多小时自主运行测试的 15 个生产级钩子。包括破坏性命令拦截器、分支保护器、语法检查器、上下文监控器和权限自动批准器。纯 Bash，无依赖 |\n| [cc-aws-keepalive](https:\u002F\u002Fgithub.com\u002FGeiserX\u002Fcc-aws-keepalive) | 通过 AWS 凭证过期保持 Claude Code 会话活跃——主动到期预警、针对 Bedrock SSO\u002FSAML 的凭证刷新绕过，以及可选的状态栏倒计时。4 个钩子脚本，无依赖 |\n| [cc-safe-setup](https:\u002F\u002Fgithub.com\u002Fyurukusa\u002Fcc-safe-setup) | 一条命令（`npx cc-safe-setup`）即可在 10 秒内安装 6 个必备的安全钩子。无依赖 |\n| [knowledge-graph](https:\u002F\u002Fgithub.com\u002Fhilyfux\u002Fknowledge-graph) | 基于 Anthropic 内部工程实践和 Karpathy 的 AutoResearch 方法论。这是一个零依赖、原生 Git 的内存层，适用于 Claude Code，能够在不同会话之间持续保留已学习的上下文。纯 Bash，每事件耗时约 3 毫秒，注重隐私 |\n| [claude-supermemory](https:\u002F\u002Fgithub.com\u002Fsupermemoryai\u002Fclaude-supermemory) | 利用 Supermemory 实现跨会话和项目的持久化记忆。在会话开始时注入用户档案，自动捕获对话内容。2,300+ 颗星 |\n| [code-architect](plugins\u002Fcode-architect\u002F) | 生成架构图和技术设计文档 |\n| [code-explainer](plugins\u002Fcode-explainer\u002F) | 解释复杂代码，并在文件中添加内联文档 |\n| [code-guardian](plugins\u002Fcode-guardian\u002F) | 自动化代码审查、安全扫描和质量保证 |\n| [code-review-assistant](plugins\u002Fcode-review-assistant\u002F) | 自动化代码审查，带有严重等级和可操作的反馈 |\n| [eliniscan](https:\u002F\u002Fgithub.com\u002FAlpYenigun\u002Feliniscan) | AI 全代码库扫描器。为每个文件开启一个独立的 Claude 会话——逐行阅读，不遗漏任何内容。能够发现 bug、安全漏洞、性能问题，并自动修复 |\n| [codebase-documenter](plugins\u002Fcodebase-documenter\u002F) | 自动为整个代码库生成内联注释和 API 文档 |\n| [color-contrast](plugins\u002Fcolor-contrast\u002F) | 颜色对比度检查和无障碍颜色建议 |\n| [commit-commands](plugins\u002Fcommit-commands\u002F) | 带有智能暂存和推送自动化的高级提交工作流 |\n| [complexity-reducer](plugins\u002Fcomplexity-reducer\u002F) | 降低圈复杂度并简化函数 |\n| [compliance-checker](plugins\u002Fcompliance-checker\u002F) | 针对 GDPR、SOC2 和 HIPAA 的合规性验证 |\n| [content-creator](plugins\u002Fcontent-creator\u002F) | 为博客文章和社交媒体生成技术内容 |\n| [context7-docs](plugins\u002Fcontext7-docs\u002F) | 通过 Context7 获取最新的库文档，以确保编码的准确性 |\n| [contract-tester](plugins\u002Fcontract-tester\u002F) | 使用 Pact 进行 API 合同测试，以确保微服务兼容性 |\n| [cozempic](https:\u002F\u002Fgithub.com\u002FRuya-AI\u002Fcozempic) | ✨ v1.2.x — 现已支持自更新、原子写入、严格的会话保护，且在团队检测方面零误报。拥有 13 种修剪策略、代理团队保护、MCP 服务器和 JSONL 医生。安装命令：`\u002Fplugin marketplace add Ruya-AI\u002Fcozempic` |\n| [create-worktrees](plugins\u002Fcreate-worktrees\u002F) | 用于并行开发工作流的 Git 工作树管理 |\n| [cron-scheduler](plugins\u002Fcron-scheduler\u002F) | Cron 作业的配置和计划验证 |\n| [css-cleaner](plugins\u002Fcss-cleaner\u002F) | 查找未使用的 CSS 并合并样式表 |\n| [cup](https:\u002F\u002Fgithub.com\u002Fkrodak\u002Fclickup-cli) | 适用于 AI 代理的 ClickUp CLI，具备任务管理、冲刺和时间追踪功能 |\n| [data-privacy](plugins\u002Fdata-privacy\u002F) | 实现数据隐私保护，包括 PII 检测和匿名化 |\n| [database-optimizer](plugins\u002Fdatabase-optimizer\u002F) | 通过索引建议和 EXPLAIN 分析优化数据库查询 |\n| [dead-code-finder](plugins\u002Fdead-code-finder\u002F) | 在整个代码库中查找并移除死代码 |\n| [debug-session](plugins\u002Fdebug-session\u002F) | 与 git bisect 集成的交互式调试工作流 |\n| [dependency-manager](plugins\u002Fdependency-manager\u002F) | 审计、更新和管理项目依赖，同时进行安全检查 |\n| [deploy-pilot](plugins\u002Fdeploy-pilot\u002F) | 自动化部署，包括 Dockerfile 生成、CI\u002FCD 流水线和基础设施即代码 |\n| [desktop-app](plugins\u002Fdesktop-app\u002F) | 使用 Electron 或 Tauri 搭建桌面应用程序框架 |\n| [devops-automator](plugins\u002Fdevops-automator\u002F) | 用于 CI\u002FCD、健康检查和部署的 DevOps 自动化脚本 |\n| [discuss](plugins\u002Fdiscuss\u002F) | 通过结构化的利弊分析来讨论实施方法 |\n| [claw-army\u002Fclaude-node](https:\u002F\u002Fgithub.com\u002Fclaw-army\u002Fclaude-node) | 用于 Claude Code CLI 的 Python 子进程桥接，使 Python 代码能够通过流式 JSON 直接访问 Claude Code 的原生功能 |\n| [discoclaw](https:\u002F\u002Fgithub.com\u002FDiscoClaw\u002Fdiscoclaw) | 个人 AI 编排器，将 Discord 与 Claude Code 桥接，具备持久记忆、任务跟踪和基于 cron 的自动化功能 |\n| [jarvis](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fjarvis) | 将闲置的 Claude Max 订阅转化为 24\u002F7 的 AI 运维系统——包含 Discord 机器人、76 个定时任务、12 个 AI 团队、本地 LanceDB RAG、通过 Nexus CIG 实现 98% 的上下文压缩，以及四层自我修复的基础设施。采用 `claude -p` 无头模式，无需额外费用。 |\n| [jarvis-company-board](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fjarvis-company-board) | 基于 Next.js 15 和 SQLite WAL 构建的实时 AI 代理协作板——8 名指定的 AI 成员通过 SSE 推送进行决策辩论，配备 DEV 任务审批工作流和 Railway 部署支持。 |\n| [dna-claude-analysis](https:\u002F\u002Fgithub.com\u002Fshmlkv\u002Fdna-claude-analysis) | 个人基因组分析工具包，可对原始 DNA 数据进行 17 类别的分析，并生成终端风格的 HTML 仪表盘 |\n| [codetape](https:\u002F\u002Fgithub.com\u002F888wing\u002Fcodetape) | AI 编码的“黑匣子”——自动记录语义轨迹，并同步 README、CHANGELOG 和 CLAUDE.md。无依赖。`npx codetape init` |\n| [doc-forge](plugins\u002Fdoc-forge\u002F) | 文档生成、API 文档和 README 的维护 |\n| [docker-helper](plugins\u002Fdocker-helper\u002F) | 构建优化的 Docker 镜像，改进 Dockerfile 最佳实践 |\n| [double-check](plugins\u002Fdouble-check\u002F) | 通过系统的二次分析验证代码正确性 |\n| [e2e-runner](plugins\u002Fe2e-runner\u002F) | 对 Web 应用程序进行端到端测试执行和记录 |\n| [embedding-manager](plugins\u002Fembedding-manager\u002F) | 管理向量嵌入和相似度搜索 |\n| [env-manager](plugins\u002Fenv-manager\u002F) | 设置和验证跨环境的环境配置 |\n| [env-sync](plugins\u002Fenv-sync\u002F) | 环境变量的同步和差异比较 |\n| [experiment-tracker](plugins\u002Fexperiment-tracker\u002F) | 使用 ML 进行实验跟踪，记录指标并比较运行结果 |\n| [explore](plugins\u002Fexplore\u002F) | 智能代码库探索，结合依赖映射和结构分析 |\n| [faf-skills](https:\u002F\u002Fgithub.com\u002FWolfe-Jam\u002Ffaf-skills) | 31 个用于维持项目上下文的 Claude Code 技能（.faf，`application\u002Fvnd.faf+yaml`）。包括评分、同步、测试、发布、MCP 服务器搭建和架构文档。Anthropic MCP #2759。 |\n| [feature-dev](plugins\u002Ffeature-dev\u002F) | 从规格说明到完成的完整功能开发流程 |\n| [fractal](https:\u002F\u002Fgithub.com\u002Frmolines\u002Ffractal) | 递归式项目管理插件。将任何目标分解为可验证的谓词，优先处理风险最高的未知部分。包含 `\u002Ffractal:run`（幂等状态机）、`\u002Ffractal:init`、`\u002Ffractal:patch`、干运行模式以及逐步分解并重新评估的功能 |\n| [finance-tracker](plugins\u002Ffinance-tracker\u002F) | 通过时间估算和预算报告跟踪开发成本 |\n| [fix-github-issue](plugins\u002Ffix-github-issue\u002F) | 通过分析问题详情并实施解决方案自动修复 GitHub 问题 |\n| [fix-pr](plugins\u002Ffix-pr\u002F) | 根据上下文自动生成补丁，自动修复 PR 审查意见 |\n| [flutter-mobile](plugins\u002Fflutter-mobile\u002F) | 使用小部件创建和平台通道进行 Flutter 应用开发 |\n| [frontend-developer](plugins\u002Ffrontend-developer\u002F) | 前端组件开发，兼顾无障碍性和响应式设计 |\n| [gcp-helper](plugins\u002Fgcp-helper\u002F) | Google Cloud Platform 服务的配置和部署 |\n| [git-flow](plugins\u002Fgit-flow\u002F) | Git 工作流管理，包括特性分支、发布和热修复流程 |\n| [github-issue-manager](plugins\u002Fgithub-issue-manager\u002F) | GitHub 问题的分类、创建和管理 |\n| [harness-evolver](https:\u002F\u002Fgithub.com\u002Fraphaelchristi\u002Fharness-evolver) | 基于 LangSmith 的自主代理优化。多代理提案者在隔离的 Git 工作树中进化提示、路由、工具和架构。安装命令：`npx harness-evolver@latest` |\n| [helm-charts](plugins\u002Fhelm-charts\u002F) | Helm 图表的生成和升级管理 |\n| [import-organizer](plugins\u002Fimport-organizer\u002F) | 整理、排序和清理导入语句 |\n| [infrastructure-maintainer](plugins\u002Finfrastructure-maintainer\u002F) | 进行基础设施维护，包括安全审计和更新管理 |\n| [ios-developer](plugins\u002Fios-developer\u002F) | 使用 SwiftUI 视图和模型进行 iOS 和 Swift 开发 |\n| [k8s-helper](plugins\u002Fk8s-helper\u002F) | 使用 kubectl 生成 Kubernetes 清单并调试 Pod 问题 |\n| [leapfrog-mcp](https:\u002F\u002Fgithub.com\u002Fanthonybono21-cloud\u002Fleapfrog) | 多会话浏览器 MCP 服务器——15 个并行隔离的 Chromium 会话，允许多个 Claude Code 终端同时浏览。包含 37 种用于导航、提取、交互、网络拦截和会话配置的工具。经过 797 次测试。安装命令：`npx -y leapfrog-mcp` |\n| [license-checker](plugins\u002Flicense-checker\u002F) | 许可证合规性检查和 NOTICE 文件生成 |\n| [lighthouse-runner](plugins\u002Flighthouse-runner\u002F) | 运行 Lighthouse 审计并修复性能问题 |\n| [lightcms](https:\u002F\u002Fgithub.com\u002Fjonradoff\u002Flightcms) | 基于 AI 的 CMS，配备 41 种 MCP 工具，可通过自然语言管理网站——页面、模板、资产、主题、集合、重定向等，并支持完整的内容版本控制 |\n| [linear-helper](plugins\u002Flinear-helper\u002F) | Linear 问题跟踪集成和工作流管理 |\n| [load-tester](plugins\u002Fload-tester\u002F) | 对 API 和 Web 服务进行负载和压力测试 |\n| [memory-profiler](plugins\u002Fmemory-profiler\u002F) | 内存泄漏检测和堆分析 |\n| [migrate-tool](plugins\u002Fmigrate-tool\u002F) | 生成数据库迁移和框架升级的代码迁移脚本 |\n| [migration-generator](plugins\u002Fmigration-generator\u002F) | 数据库迁移的生成和回滚管理 |\n| [model-context-protocol](plugins\u002Fmodel-context-protocol\u002F) | 用于 MCP 服务器开发的帮助工具，提供工具和资源脚手架 |\n| [model-evaluator](plugins\u002Fmodel-evaluator\u002F) | 评估和比较 ML 模型的性能指标 |\n| [monitoring-setup](plugins\u002Fmonitoring-setup\u002F) | 监控和警报配置，包括仪表盘生成 |\n| [monorepo-manager](plugins\u002Fmonorepo-manager\u002F) | 管理 monorepo 包，检测受影响的包并同步版本 |\n| [mutation-tester](plugins\u002Fmutation-tester\u002F) | 通过突变测试衡量测试套件的质量 |\n| [myclaude](https:\u002F\u002Fgithub.com\u002Fstellarlinkco\u002Fmyclaude) | 多代理编排系统，根据复杂程度将任务分配给 Claude Code、Codex、Gemini 和 OpenCode。OmO 技能用于智能路由。2,400+ 颗星 |\n| [nimbalyst](https:\u002F\u002Fnimbalyst.com) | 用于使用 Codex 和 Claude Code 进行构建的可视化工作区。会话和任务管理器。可视化编辑 |\n| [n8n-workflow](plugins\u002Fn8n-workflow\u002F) | 根据自然语言描述生成 n8n 自动化工作流 |\n| [onboarding-guide](plugins\u002Fonboarding-guide\u002F) | 新开发者入职文档生成器 |\n| [oh-my-claudecode](https:\u002F\u002Fgithub.com\u002FYeachan-Heo\u002Foh-my-claudecode) | 以团队为核心的多代理编排。包含 19 个专业代理和 28 个技能。完全自主执行，采用苏格拉底式提问，协调 9 个代理。9,900+ 颗星 |\n| [onWatch](https:\u002F\u002Fgithub.com\u002Fonllm-dev\u002Fonwatch) | 开源 Go CLI，用于跟踪 7 家提供商（Synthetic、Z.ai、Anthropic、Codex、GitHub Copilot、MiniMax、Antigravity）的 AI API 配额使用情况。后台守护进程占用不到 50 MB 内存，无遥测数据，并配有 Material Design 3 风格的 Web 仪表盘 |\n| [obey](https:\u002F\u002Fgithub.com\u002FLexxes-Projects\u002Fobey) | 规则执行插件。以自然语言保存规则，通过 17 个生命周期钩子强制执行。包含三个作用范围（全局、栈特定、项目本地），通过 PreToolUse 实现主动阻止，通过 Stop 钩子进行完成检查清单，具备审计追踪功能，并能自动检测类似规则的语言。跨平台适用。 |\n| [opcode](https:\u002F\u002Fgithub.com\u002Fwinfunc\u002Fopcode) | 用于 Claude Code 的 Tauri 2 桌面 GUI 和工具包。管理会话、使用可视化编辑器创建自定义代理、查看使用情况分析，并集成 MCP。21,000+ 颗星 |\n| [openapi-expert](plugins\u002Fopenapi-expert\u002F) | OpenAPI 规范的生成、验证和客户端代码脚手架 |\n| [optimize](plugins\u002Foptimize\u002F) | 代码优化，以提升性能并减小包体积 |\n| [ORCH](https:\u002F\u002Fgithub.com\u002Foxgeneral\u002FORCH) | 一个 CLI 运行时，将 Claude Code、Codex 和 Cursor 编排为类型化的代理团队，配备状态机（待办→审核→完成）、自动重试、代理间消息传递、目标以及 TUI 仪表盘 |\n| [oss-autopilot](https:\u002F\u002Fgithub.com\u002Fcostajohnt\u002Foss-autopilot) | 开源贡献管理者——跟踪跨仓库的 PR，发现问题，诊断 CI 失败，并起草维护人员的回复。包含 7 个代理、交互式命令和 MCP 服务器。安装命令：`\u002Fplugin marketplace add costajohnt\u002Foss-autopilot` |\n| [paco-framework](https:\u002F\u002Fgithub.com\u002FPenguinAlleyApps\u002Fpaco-framework) | 1+ | 以 Markdown 优先的多代理操作系统，专为 Claude Code 设计。协调 3–16 个专业 AI 代理（工程、QA、增长、财务等），通过基于文件的调度、机构记忆、CEO Gate 批准以及 24\u002F7 排程运作。无需 Python。MIT 许可证 |\n| [peon-ping](https:\u002F\u002Fgithub.com\u002FPeonPing\u002Fpeon-ping) | 为 Claude Code 和其他代理提供《魔兽争霸 III》Peon 语音通知（还包括《星际争霸》、《传送门》和《塞尔达传说》）。桌面横幅，可自动检测 SSH\u002FDevcontainers。3,900+ 颗星 |\n| [perf-profiler](plugins\u002Fperf-profiler\u002F) | 性能分析、剖析和优化建议 |\n| [performance-monitor](plugins\u002Fperformance-monitor\u002F) | 对 API 端点进行 profiling，并运行基准测试以识别性能瓶颈 |\n| [plan](plugins\u002Fplan\u002F) | 结构化规划，包含风险评估和时间估算 |\n| [preflight](https:\u002F\u002Fgithub.com\u002Fpreflight-dev\u002Fpreflight) | 一个包含 24 种工具的 MCP 服务器，可在模糊提示导致错误→修复循环之前将其捕捉。配备 12 类别评分卡、纠正模式学习、跨服务合同意识、使用 LanceDB 向量搜索会话历史，以及成本估算功能。安装命令：`npx preflight-dev` |\n| [pr-reviewer](plugins\u002Fpr-reviewer\u002F) | 通过结构化分析审查拉取请求，并自信地批准 |\n| [product-shipper](plugins\u002Fproduct-shipper\u002F) | 从头到尾交付功能，包括上线检查清单和推广计划 |\n| [production-grade](https:\u002F\u002Fgithub.com\u002Fnagisanzenin\u002Fclaude-code-production-grade-plugin) | 一个由 14 个代理组成的自主流水线——项目经理、架构师、后端、前端、QA、安全、代码审查、DevOps、SRE、数据科学家、技术作家、技能创造者和博学的副驾驶。两波并行执行，对现有系统安全。 |\n| [project-scaffold](plugins\u002Fproject-scaffold\u002F) | 使用最佳实践模板搭建新项目并添加功能 |\n| [prompt-optimizer](plugins\u002Fprompt-optimizer\u002F) | 分析并优化 AI 提示，以获得更好的效果 |\n| [pulse](https:\u002F\u002Fgithub.com\u002Fchsm04\u002Fpulse) | 本地频道插件——通过 HTTP POST 将通知推送到 Claude Code 会话中。无需 Discord\u002FSlack，只需 curl 命令。包含三个级别（信息\u002F警告\u002F错误），具备来源追踪和去重功能。 |\n| [PUIUX Pilot](https:\u002F\u002Fgithub.com\u002FPUIUX-Cloud\u002Fpuiux-pilot) | 自动为任何项目配置 Claude Code 钩子、MCP 和技能。扫描 95 种以上项目类型，从 28 种以上钩子中选择，给出 0–100 分的质量评分，并在不同 AI 工具之间转换配置。安全措施包括干运行、原子写入、备份和回滚。 |\n| [python-expert](plugins\u002Fpython-expert\u002F) | 针对 Python 的开发，包括类型提示和惯用重构 |\n| [pulser](https:\u002F\u002Fgithub.com\u002Fwhynowlab\u002Fpulser) | 用于 Claude Code SKILL.md 文件的诊断 CLI——扫描 8 条规则，自动分类，并提供带有模板的修复建议，支持带回滚的 --fix 参数 |\n| [query-optimizer](plugins\u002Fquery-optimizer\u002F) | SQL 查询优化和执行计划分析 |\n| [rag-builder](plugins\u002Frag-builder\u002F) | 构建检索增强生成管道 |\n| [rapid-prototyper](plugins\u002Frapid-prototyper\u002F) | 快速搭建原型，仅需最简结构 |\n| [reporecall](https:\u002F\u002Fgithub.com\u002Fproofofwork-agency\u002Freporecall) | 为 Claude Code 提供本地代码库记忆。采用 Tree-sitter AST 索引（支持 22 种语言），混合关键词和向量搜索，遍历调用图，并结合钩子和 MCP。在 Claude 开始思考前约 5 毫秒注入相关上下文。npm：`@proofofwork-agency\u002Freporecall` |\n| [react-native-dev](plugins\u002Freact-native-dev\u002F) | 使用特定于平台的优化进行 React Native 移动端开发 |\n| [readme-generator](plugins\u002Freadme-generator\u002F) | 根据项目分析智能生成 README |\n| [refactor-engine](plugins\u002Frefactor-engine\u002F) | 提取函数、简化复杂代码，并降低认知复杂度 |\n| [regex-builder](plugins\u002Fregex-builder\u002F) | 构建、测试和调试正则表达式模式 |\n| [release-manager](plugins\u002Frelease-manager\u002F) | 语义版本号管理和自动化发布工作流 |\n| [responsive-designer](plugins\u002Fresponsive-designer\u002F) | 实现响应式设计并进行测试 |\n| [review-squad](https:\u002F\u002Fgithub.com\u002F2389-research\u002Freview-squad) | 通过派遣子代理小组进行多视角代码审查：专家、普通用户、吹毛求疵者以及真实用户任务执行者 |\n| [schema-designer](plugins\u002Fschema-designer\u002F) | 数据库模式设计和 ERD 生成 |\n| [screen-reader-tester](plugins\u002Fscreen-reader-tester\u002F) | 测试屏幕阅读器兼容性并修复 ARIA 属性 |\n| [security-guidance](plugins\u002Fsecurity-guidance\u002F) | 安全最佳实践顾问，提供漏洞检测和修复建议 |\n| [seed-generator](plugins\u002Fseed-generator\u002F) | 生成包含真实数据的数据库种子脚本 |\n| [sitemd](https:\u002F\u002Fgithub.com\u002Fsitemd-cc\u002Fsitemd) | 使用 MCP 从 Markdown 构建网站。包含 22 种工具，用于创建页面、生成内容、配置设置、进行 SEO 审计以及将静态站点部署到 Cloudflare Pages |\n| [simmer](https:\u002F\u002Fgithub.com\u002F2389-research\u002Fsimmer) | 通过法官子代理对成果进行迭代优化，这些子代理会根据用户定义的标准在多轮中进行评分 |\n| [slack-notifier](plugins\u002Fslack-notifier\u002F) | Slack 集成，用于部署和构建通知 |\n| [smart-commit](plugins\u002Fsmart-commit\u002F) | 智能 Git 提交，格式规范、语义分析并生成变更日志 |\n| [sprint-prioritizer](plugins\u002Fsprint-prioritizer\u002F) | 根据故事优先级和容量估算进行 Sprint 计划 |\n| [technical-sales](plugins\u002Ftechnical-sales\u002F) | 制作技术演示并撰写 PoC 方案 |\n| [the-pragmatic-pm](https:\u002F\u002Fgithub.com\u002Fmarfoerst\u002Fthe-pragmatic-pm) | PM 领导力工具包，包含 43 个技能、5 个代理和 4 个工作流。涵盖 PRD 生成、OKR 生命周期、定价、AI 定价、定位、销售赋能以及季度计划制定。 |\n| [terraform-helper](plugins\u002Fterraform-helper\u002F) | 创建 Terraform 模块和进行基础设施规划 |\n| [test-data-generator](plugins\u002Ftest-data-generator\u002F) | 生成逼真的测试数据并填充数据库 |\n| [test-kitchen](https:\u002F\u002Fgithub.com\u002F2389-research\u002Ftest-kitchen) | 使用竞争性的子代理进行并行实施探索，通过结构化比较选出优胜者 |\n| [test-results-analyzer](plugins\u002Ftest-results-analyzer\u002F) | 分析测试失败，识别模式，并提出针对性的修复建议 |\n| [test-writer](plugins\u002Ftest-writer\u002F) | 生成全面的单元和集成测试，覆盖所有代码路径 |\n| [tool-evaluator](plugins\u002Ftool-evaluator\u002F) | 根据结构化的评分标准评估和比较开发者工具 |\n| [type-migrator](plugins\u002Ftype-migrator\u002F) | 将 JavaScript 文件迁移到 TypeScript，赋予其正确的类型 |\n| [ui-designer](plugins\u002Fui-designer\u002F) | 根据规格书实现 UI 设计，生成像素级精确的组件 |\n| [ultrathink](plugins\u002Fultrathink\u002F) | 深度分析模式，针对复杂问题提供更长时间的推理 |\n| [unit-test-generator](plugins\u002Funit-test-generator\u002F) | 为任何函数或模块生成全面的单元测试 |\n| [update-branch](plugins\u002Fupdate-branch\u002F) | 通过解决冲突重新基线和更新特性分支 |\n| [vision-specialist](plugins\u002Fvision-specialist\u002F) | 图像和视觉分析，包括截图解读和文本提取 |\n| [vibe-kanban](https:\u002F\u002Fgithub.com\u002FBloopAI\u002Fvibe-kanban) | 基于 Kanban 的编排，适用于 10 多个编码代理（Claude Code、Codex、Gemini CLI、Copilot 等）。每个代理使用独立的 Git 工作树，进行内联差异审查。23,200+ 颗星 |\n| [VibeGuard](https:\u002F\u002Fgithub.com\u002Fmajiayu000\u002Fvibeguard) | 阻止 AI 幻想代码。包含 88 条规则、13 个钩子和 14 个代理，用于实时拦截和静态扫描，覆盖 5 种语言 |\n| [visual-regression](plugins\u002Fvisual-regression\u002F) | 使用截图比较进行视觉回归测试 |\n| [wshobson\u002Fagents](https:\u002F\u002Fgithub.com\u002Fwshobson\u002Fagents) | 包括 112 个专业代理、16 个多代理工作流编排者、146 个技能以及 79 种工具，分布在 72 个专注插件中。31,300+ 颗星 |\n| [web-dev](plugins\u002Fweb-dev\u002F) | 全栈 Web 开发，包括应用脚手架和页面生成 |\n| [workflow-optimizer](plugins\u002Fworkflow-optimizer\u002F) | 对开发工作流进行分析并提出优化建议 |\n| [background-timer](https:\u002F\u002Fgithub.com\u002FculminationAI\u002Fbackground-timer) | 背景计时器，带有任务通知功能——可在不阻塞对话的情况下设置延迟检查 |\n| [claude-sounds](https:\u002F\u002Fgithub.com\u002FculminationAI\u002Fclaude-sounds) | 为 Claude Code 钩子提供音频反馈——10 种事件、21 种声音，随机轮换，可自定义（适用于 macOS）|\n\n| [notch-so-good](https:\u002F\u002Fgithub.com\u002Fdeepshal99\u002Fnotch-so-good) | 一款像素风螃蟹（Chawd）住在你的Mac屏幕刘海里，帮你盯着Claude Code。支持实时会话计时器、颜色编码通知、13种空闲动画、鼠标互动的眼睛以及困倦提醒系统。通用二进制文件，一行命令即可安装：`npx notch-so-good`。采用MIT许可证，拥有130多位用户 |\n| [codebase-graph](https:\u002F\u002Fgithub.com\u002FPhoenixrr2113\u002Fcodebase-graph) | 代码智能MCP服务器——支持42种语言的tree-sitter AST解析、FalkorDB知识图谱，搜索质量达到0.944 MRR。可通过npm安装：`@anthropic\u002Fcodegraph` |\n\n\n\n### 安装插件\n\n```bash\n\u002Fplugin install claude-code-toolkit@smart-commit\n```\n\n或者一次性安装所有插件：\n\n```bash\n\u002Fplugin install claude-code-toolkit\n```\n\n---\n\n## 代理\n\n一百三十五个专业代理被划分为十个类别。每个代理都定义了角色定位、系统指令以及工具使用模式。\n\n### 核心开发类（13个代理）\n\n| 代理 | 文件 | 用途 |\n|-------|------|---------|\n| 全栈工程师 | [`fullstack-engineer.md`](agents\u002Fcore-development\u002Ffullstack-engineer.md) | 负责从前端到后端再到数据库的端到端功能交付 |\n| API设计师 | [`api-designer.md`](agents\u002Fcore-development\u002Fapi-designer.md) | 使用OpenAPI设计RESTful API，实现版本控制与分页功能 |\n| 前端架构师 | [`frontend-architect.md`](agents\u002Fcore-development\u002Ffrontend-architect.md) | 负责组件架构、状态管理及性能优化 |\n| 移动开发者 | [`mobile-developer.md`](agents\u002Fcore-development\u002Fmobile-developer.md) | 使用React Native和Flutter进行跨平台移动应用开发 |\n| 后端开发者 | [`backend-developer.md`](agents\u002Fcore-development\u002Fbackend-developer.md) | 构建Node.js\u002FExpress\u002FFastify后端服务 |\n| GraphQL架构师 | [`graphql-architect.md`](agents\u002Fcore-development\u002Fgraphql-architect.md) | 负责Schema设计、Resolver、联邦架构及DataLoader实现 |\n| 微服务架构师 | [`microservices-architect.md`](agents\u002Fcore-development\u002Fmicroservices-architect.md) | 设计分布式系统、事件驱动架构及Saga模式 |\n| WebSocket工程师 | [`websocket-engineer.md`](agents\u002Fcore-development\u002Fwebsocket-engineer.md) | 实现实时通信，使用Socket.io并进行扩展 |\n| UI设计师 | [`ui-designer.md`](agents\u002Fcore-development\u002Fui-designer.md) | 负责UI\u002FUX实现、设计系统搭建以及从Figma转码为代码 |\n| Electron开发者 | [`electron-developer.md`](agents\u002Fcore-development\u002Felectron-developer.md) | 开发Electron桌面应用，处理IPC通信及原生操作系统集成 |\n| API网关工程师 | [`api-gateway-engineer.md`](agents\u002Fcore-development\u002Fapi-gateway-engineer.md) | 设计API网关模式，实现限流与认证代理功能 |\n| 单体仓库架构师 | [`monorepo-architect.md`](agents\u002Fcore-development\u002Fmonorepo-architect.md) | 制定Turborepo\u002FNx工作区策略，管理依赖关系图 |\n| 事件驱动架构师 | [`event-driven-architect.md`](agents\u002Fcore-development\u002Fevent-driven-architect.md) | 负责事件溯源、CQRS、消息队列及分布式事件的设计 |\n\n### 语言专家类（25个代理）\n\n| 代理 | 文件 | 用途 |\n|-------|------|---------|\n| TypeScript专家 | [`typescript-specialist.md`](agents\u002Flanguage-experts\u002Ftypescript-specialist.md) | 负责类型安全模式、泛型及模块设计 |\n| Python工程师 | [`python-engineer.md`](agents\u002Flanguage-experts\u002Fpython-engineer.md) | 掌握Pythonic风格编程、打包及异步操作 |\n| Rust系统工程师 | [`rust-systems.md`](agents\u002Flanguage-experts\u002Frust-systems.md) | 负责所有权、生命周期及Trait设计 |\n| Go开发者 | [`golang-developer.md`](agents\u002Flanguage-experts\u002Fgolang-developer.md) | 熟练使用接口、goroutine及错误处理机制 |\n| Next.js开发者 | [`nextjs-developer.md`](agents\u002Flanguage-experts\u002Fnextjs-developer.md) | 负责App Router、RSC、ISR及服务器端动作的实现 |\n| React专家 | [`react-specialist.md`](agents\u002Flanguage-experts\u002Freact-specialist.md) | 精通React 19新特性、Hooks及状态管理 |\n| Django开发者 | [`django-developer.md`](agents\u002Flanguage-experts\u002Fdjango-developer.md) | 掌握Django 5+版本的新特性、DRF框架及ORM优化技巧 |\n| Rails专家 | [`rails-expert.md`](agents\u002Flanguage-experts\u002Frails-expert.md) | 熟悉Rails 7+版本的新特性，包括Hotwire和ActiveRecord |\n| Java架构师 | [`java-architect.md`](agents\u002Flanguage-experts\u002Fjava-architect.md) | 负责Spring Boot 3+版本的应用开发，以及JPA和微服务相关技术 |\n| Kotlin专家 | [`kotlin-specialist.md`](agents\u002Flanguage-experts\u002Fkotlin-specialist.md) | 熟练使用协程、Ktor框架及多平台开发 |\n| Flutter专家 | [`flutter-expert.md`](agents\u002Flanguage-experts\u002Fflutter-expert.md) | 掌握Flutter 3+版本的新特性、Dart语言及Riverpod状态管理库 |\n| C#开发者 | [`csharp-developer.md`](agents\u002Flanguage-experts\u002Fcsharp-developer.md) | 熟悉.NET 8+版本、ASP.NET Core及EF Core框架 |\n| PHP开发者 | [`php-developer.md`](agents\u002Flanguage-experts\u002Fphp-developer.md) | 掌握PHP 8.3+版本的新特性、Laravel 11框架及Eloquent ORM |\n| Elixir专家 | [`elixir-expert.md`](agents\u002Flanguage-experts\u002Felixir-expert.md) | 熟悉OTP架构、Phoenix LiveView及Ecto数据库操作 |\n| Angular架构师 | [`angular-architect.md`](agents\u002Flanguage-experts\u002Fangular-architect.md) | 掌握Angular 17+版本的新特性，包括signals和独立组件 |\n| Vue专家 | [`vue-specialist.md`](agents\u002Flanguage-experts\u002Fvue-specialist.md) | 熟悉Vue 3版本的新特性、Composition API、Pinia状态管理及Nuxt框架 |\n| Svelte开发者 | [`svelte-developer.md`](agents\u002Flanguage-experts\u002Fsvelte-developer.md) | 负责SvelteKit框架、Runes语法及表单操作的实现 |\n| Swift开发者 | [`swift-developer.md`](agents\u002Flanguage-experts\u002Fswift-developer.md) | 熟练使用SwiftUI、iOS 17+版本的新特性、Combine框架及结构化并发技术 |\n| Scala开发者 | [`scala-developer.md`](agents\u002Flanguage-experts\u002Fscala-developer.md) | 熟悉Akka Actors、Play Framework及Cats Effect库 |\n| Haskell开发者 | [`haskell-developer.md`](agents\u002Flanguage-experts\u002Fhaskell-developer.md) | 纯函数式编程、Monad、Type Class及GHC扩展的使用 |\n| Lua开发者 | [`lua-developer.md`](agents\u002Flanguage-experts\u002Flua-developer.md) | 负责游戏脚本编写、Neovim插件开发及LuaJIT加速引擎的使用 |\n| Zig开发者 | [`zig-developer.md`](agents\u002Flanguage-experts\u002Fzig-developer.md) | 系统级编程、编译期计算及内存分配策略的掌握 |\n| Clojure开发者 | [`clojure-developer.md`](agents\u002Flanguage-experts\u002Fclojure-developer.md) | 熟悉REPL驱动开发、Ring\u002FCompojure框架及ClojureScript语言 |\n| OCaml开发者 | [`ocaml-developer.md`](agents\u002Flanguage-experts\u002Focaml-developer.md) | 熟练使用类型推断、模式匹配及Dream框架 |\n| Nim开发者 | [`nim-developer.md`](agents\u002Flanguage-experts\u002Fnim-developer.md) | 熟悉元编程、GC策略及与C\u002FC++的互操作性 |\n\n### 基础设施（11个角色）\n\n| 角色 | 文件 | 用途 |\n|-------|------|---------|\n| 云架构师 | [`cloud-architect.md`](agents\u002Finfrastructure\u002Fcloud-architect.md) | AWS、GCP、Azure 的资源 provision 和 IaC |\n| DevOps 工程师 | [`devops-engineer.md`](agents\u002Finfrastructure\u002Fdevops-engineer.md) | CI\u002FCD、容器化、监控 |\n| 数据库管理员 | [`database-admin.md`](agents\u002Finfrastructure\u002Fdatabase-admin.md) | 模式设计、查询调优、复制 |\n| 平台工程师 | [`platform-engineer.md`](agents\u002Finfrastructure\u002Fplatform-engineer.md) | 内部开发者平台、服务目录 |\n| Kubernetes 专家 | [`kubernetes-specialist.md`](agents\u002Finfrastructure\u002Fkubernetes-specialist.md) | Operators、CRD、服务网格、Istio |\n| Terraform 工程师 | [`terraform-engineer.md`](agents\u002Finfrastructure\u002Fterraform-engineer.md) | IaC、模块设计、状态管理、多云部署 |\n| 网络工程师 | [`network-engineer.md`](agents\u002Finfrastructure\u002Fnetwork-engineer.md) | DNS、负载均衡器、CDN、防火墙规则 |\n| SRE 工程师 | [`sre-engineer.md`](agents\u002Finfrastructure\u002Fsre-engineer.md) | SLO、错误预算、事件响应、事后分析 |\n| 部署工程师 | [`deployment-engineer.md`](agents\u002Finfrastructure\u002Fdeployment-engineer.md) | 蓝绿部署、金丝雀发布、滚动更新 |\n| 安全工程师 | [`security-engineer.md`](agents\u002Finfrastructure\u002Fsecurity-engineer.md) | IAM 策略、mTLS、秘钥管理、Vault |\n| 事件响应者 | [`incident-responder.md`](agents\u002Finfrastructure\u002Fincident-responder.md) | 事件分类、运行手册、沟通、恢复 |\n\n### 质量保证（10个角色）\n\n| 角色 | 文件 | 用途 |\n|-------|------|---------|\n| 代码审查员 | [`code-reviewer.md`](agents\u002Fquality-assurance\u002Fcode-reviewer.md) | 以安全和性能为重点的 PR 审查 |\n| 测试架构师 | [`test-architect.md`](agents\u002Fquality-assurance\u002Ftest-architect.md) | 测试策略、金字塔模型、覆盖率目标 |\n| 安全审计员 | [`security-auditor.md`](agents\u002Fquality-assurance\u002Fsecurity-auditor.md) | 漏洞扫描、OWASP 合规性检查 |\n| 性能工程师 | [`performance-engineer.md`](agents\u002Fquality-assurance\u002Fperformance-engineer.md) | 负载测试、性能剖析、优化 |\n| 无障碍专家 | [`accessibility-specialist.md`](agents\u002Fquality-assurance\u002Faccessibility-specialist.md) | WCAG 合规、ARIA 属性、屏幕阅读器支持 |\n| 混沌工程 | [`chaos-engineer.md`](agents\u002Fquality-assurance\u002Fchaos-engineer.md) | 混沌测试、故障注入、韧性验证 |\n| 渗透测试员 | [`penetration-tester.md`](agents\u002Fquality-assurance\u002Fpenetration-tester.md) | OWASP Top 10 漏洞评估、漏洞报告 |\n| QA 自动化 | [`qa-automation.md`](agents\u002Fquality-assurance\u002Fqa-automation.md) | 测试自动化框架、CI 集成 |\n| 合规审计员 | [`compliance-auditor.md`](agents\u002Fquality-assurance\u002Fcompliance-auditor.md) | SOC 2、GDPR、HIPAA 合规性检查 |\n| 错误检测员 | [`error-detective.md`](agents\u002Fquality-assurance\u002Ferror-detective.md) | 错误跟踪、堆栈信息分析、根因定位 |\n\n### 数据与 AI（15个角色）\n\n| 角色 | 文件 | 用途 |\n|-------|------|---------|\n| AI 工程师 | [`ai-engineer.md`](agents\u002Fdata-ai\u002Fai-engineer.md) | AI 应用集成、RAG、智能体 |\n| 机器学习工程师 | [`ml-engineer.md`](agents\u002Fdata-ai\u002Fml-engineer.md) | 机器学习流水线、训练、评估 |\n| 数据科学家 | [`data-scientist.md`](agents\u002Fdata-ai\u002Fdata-scientist.md) | 统计分析、数据可视化 |\n| 数据工程师 | [`data-engineer.md`](agents\u002Fdata-ai\u002Fdata-engineer.md) | ETL 流水线、Spark、数据仓库 |\n| 大语言模型架构师 | [`llm-architect.md`](agents\u002Fdata-ai\u002Fllm-architect.md) | 微调、模型选择、模型推理部署 |\n| 提示词工程师 | [`prompt-engineer.md`](agents\u002Fdata-ai\u002Fprompt-engineer.md) | 提示词优化、结构化输出生成 |\n| MLOps 工程师 | [`mlops-engineer.md`](agents\u002Fdata-ai\u002Fmlops-engineer.md) | 模型部署、监控、A\u002FB 测试 |\n| 自然语言处理工程师 | [`nlp-engineer.md`](agents\u002Fdata-ai\u002Fnlp-engineer.md) | NLP 流水线、词嵌入、文本分类 |\n| 数据库优化师 | [`database-optimizer.md`](agents\u002Fdata-ai\u002Fdatabase-optimizer.md) | 查询优化、索引、分区 |\n| 计算机视觉工程师 | [`computer-vision-engineer.md`](agents\u002Fdata-ai\u002Fcomputer-vision-engineer.md) | 图像分类、目标检测、PyTorch 框架 |\n| 推荐系统工程师 | [`recommendation-engine.md`](agents\u002Fdata-ai\u002Frecommendation-engine.md) | 协同过滤、基于内容、混合推荐算法 |\n| ETL 专家 | [`etl-specialist.md`](agents\u002Fdata-ai\u002Fetl-specialist.md) | 数据管道、模式演进、数据质量 |\n| 向量数据库工程师 | [`vector-database-engineer.md`](agents\u002Fdata-ai\u002Fvector-database-engineer.md) | FAISS、Pinecone、Qdrant、Weaviate、向量存储与检索 |\n| 数据可视化工程师 | [`data-visualization.md`](agents\u002Fdata-ai\u002Fdata-visualization.md) | D3.js、Chart.js、Matplotlib、Plotly 仪表盘 |\n| 特征工程 | [`feature-engineer.md`](agents\u002Fdata-ai\u002Ffeature-engineer.md) | 特征存储、特征流水线、编码策略 |\n| 自动研究代理 | [`autoresearch-agent.md`](agents\u002Fdata-ai\u002Fautoresearch-agent.md) | 基于树搜索的机器学习实验自动化、代码优化 |\n\n### 开发者体验（15个智能体）\n\n| 智能体 | 文件 | 用途 |\n|-------|------|---------|\n| CLI 开发者 | [`cli-developer.md`](agents\u002Fdeveloper-experience\u002Fcli-developer.md) | 使用 Commander、yargs、clap 的 CLI 工具 |\n| DX 优化器 | [`dx-optimizer.md`](agents\u002Fdeveloper-experience\u002Fdx-optimizer.md) | 开发者体验、工具链、人体工学 |\n| 文档工程师 | [`documentation-engineer.md`](agents\u002Fdeveloper-experience\u002Fdocumentation-engineer.md) | 技术写作、API 文档、指南 |\n| 构建工程师 | [`build-engineer.md`](agents\u002Fdeveloper-experience\u002Fbuild-engineer.md) | 构建系统、打包工具、编译 |\n| 依赖管理器 | [`dependency-manager.md`](agents\u002Fdeveloper-experience\u002Fdependency-manager.md) | 依赖审计、更新、锁定文件 |\n| 重构专家 | [`refactoring-specialist.md`](agents\u002Fdeveloper-experience\u002Frefactoring-specialist.md) | 代码重构、移除死代码 |\n| 遗留系统现代化者 | [`legacy-modernizer.md`](agents\u002Fdeveloper-experience\u002Flegacy-modernizer.md) | 遗留代码库迁移策略 |\n| MCP 开发者 | [`mcp-developer.md`](agents\u002Fdeveloper-experience\u002Fmcp-developer.md) | MCP 服务器及工具开发 |\n| 工具链工程师 | [`tooling-engineer.md`](agents\u002Fdeveloper-experience\u002Ftooling-engineer.md) | ESLint、Prettier、自定义工具链 |\n| Git 工作流管理者 | [`git-workflow-manager.md`](agents\u002Fdeveloper-experience\u002Fgit-workflow-manager.md) | 分支策略、CI、CODEOWNERS |\n| API 文档 | [`api-documentation.md`](agents\u002Fdeveloper-experience\u002Fapi-documentation.md) | OpenAPI\u002FSwagger、Redoc、交互式示例 |\n| 单仓库工具链 | [`monorepo-tooling.md`](agents\u002Fdeveloper-experience\u002Fmonorepo-tooling.md) | Changesets、工作区依赖、版本管理 |\n| VS Code 扩展 | [`vscode-extension.md`](agents\u002Fdeveloper-experience\u002Fvscode-extension.md) | LSP 集成、自定义编辑器、Webview 面板 |\n| 测试基础设施 | [`testing-infrastructure.md`](agents\u002Fdeveloper-experience\u002Ftesting-infrastructure.md) | 测试运行器、CI 分割、不稳定测试管理 |\n| 开发者门户 | [`developer-portal.md`](agents\u002Fdeveloper-experience\u002Fdeveloper-portal.md) | Backstage、服务目录、自助式基础设施 |\n\n### 专业领域（15个智能体）\n\n| 智能体 | 文件 | 用途 |\n|-------|------|---------|\n| 区块链开发者 | [`blockchain-developer.md`](agents\u002Fspecialized-domains\u002Fblockchain-developer.md) | 智能合约、Solidity、Web3 |\n| 游戏开发者 | [`game-developer.md`](agents\u002Fspecialized-domains\u002Fgame-developer.md) | 游戏逻辑、ECS、状态机 |\n| 嵌入式系统 | [`embedded-systems.md`](agents\u002Fspecialized-domains\u002Fembedded-systems.md) | 固件、RTOS、硬件接口 |\n| 金融科技工程师 | [`fintech-engineer.md`](agents\u002Fspecialized-domains\u002Ffintech-engineer.md) | 金融系统、合规性、精度要求 |\n| 物联网工程师 | [`iot-engineer.md`](agents\u002Fspecialized-domains\u002Fiot-engineer.md) | MQTT、边缘计算、数字孪生 |\n| 支付集成 | [`payment-integration.md`](agents\u002Fspecialized-domains\u002Fpayment-integration.md) | Stripe、PCI DSS、3D Secure |\n| SEO 专家 | [`seo-specialist.md`](agents\u002Fspecialized-domains\u002Fseo-specialist.md) | 结构化数据、Core Web Vitals |\n| 电商工程师 | [`e-commerce-engineer.md`](agents\u002Fspecialized-domains\u002Fe-commerce-engineer.md) | 购物车、库存、订单管理 |\n| 医疗健康工程师 | [`healthcare-engineer.md`](agents\u002Fspecialized-domains\u002Fhealthcare-engineer.md) | HIPAA、HL7 FHIR、医疗数据管道 |\n| 房地产科技 | [`real-estate-tech.md`](agents\u002Fspecialized-domains\u002Freal-estate-tech.md) | MLS 集成、地理空间搜索、估值 |\n| 教育科技 | [`education-tech.md`](agents\u002Fspecialized-domains\u002Feducation-tech.md) | LMS、SCORM\u002FxAPI、适应性学习、评估 |\n| 媒体流媒体 | [`media-streaming.md`](agents\u002Fspecialized-domains\u002Fmedia-streaming.md) | HLS\u002FDASH、转码、CDN、自适应码率 |\n| 地理空间工程师 | [`geospatial-engineer.md`](agents\u002Fspecialized-domains\u002Fgeospatial-engineer.md) | PostGIS、空间查询、地图 API、瓦片 |\n| 机器人工程师 | [`robotics-engineer.md`](agents\u002Fspecialized-domains\u002Frobotics-engineer.md) | ROS2、传感器融合、运动规划、SLAM |\n| 语音助手 | [`voice-assistant.md`](agents\u002Fspecialized-domains\u002Fvoice-assistant.md) | STT、TTS、对话管理、Alexa\u002FGoogle |\n\n### 商业与产品（12个智能体）\n\n| 智能体 | 文件 | 用途 |\n|-------|------|---------|\n| 产品经理 | [`product-manager.md`](agents\u002Fbusiness-product\u002Fproduct-manager.md) | PRD、用户故事、RICE 优先级排序 |\n| 技术文档撰写者 | [`technical-writer.md`](agents\u002Fbusiness-product\u002Ftechnical-writer.md) | 文档编写、风格指南 |\n| UX 研究员 | [`ux-researcher.md`](agents\u002Fbusiness-product\u002Fux-researcher.md) | 易用性测试、问卷设计 |\n| 项目经理 | [`project-manager.md`](agents\u002Fbusiness-product\u002Fproject-manager.md) | 冲刺计划、敏捷开发、任务跟踪 |\n| Scrum 主管 | [`scrum-master.md`](agents\u002Fbusiness-product\u002Fscrum-master.md) | 仪式、速度、回顾会议 |\n| 商业分析师 | [`business-analyst.md`](agents\u002Fbusiness-product\u002Fbusiness-analyst.md) | 需求分析、流程映射 |\n| 内容策略师 | [`content-strategist.md`](agents\u002Fbusiness-product\u002Fcontent-strategist.md) | SEO 内容、编辑日历、主题聚类 |\n| 增长工程师 | [`growth-engineer.md`](agents\u002Fbusiness-product\u002Fgrowth-engineer.md) | A\u002FB 测试、数据分析、漏斗优化 |\n| 客户成功 | [`customer-success.md`](agents\u002Fbusiness-product\u002Fcustomer-success.md) | 工单分类、知识库、健康评分 |\n| 销售工程师 | [`sales-engineer.md`](agents\u002Fbusiness-product\u002Fsales-engineer.md) | 技术演示、POC、集成指南 |\n| 法律顾问 | [`legal-advisor.md`](agents\u002Fbusiness-product\u002Flegal-advisor.md) | 使用条款、隐私政策、软件许可 |\n| 市场分析员 | [`marketing-analyst.md`](agents\u002Fbusiness-product\u002Fmarketing-analyst.md) | 活动分析、归因分析、ROI 跟踪 |\n\n### 编排（8个智能体）\n\n| 智能体 | 文件 | 用途 |\n|-------|------|---------|\n| 任务协调员 | [`task-coordinator.md`](agents\u002Forchestration\u002Ftask-coordinator.md) | 在各智能体间路由任务，管理交接工作 |\n| 上下文管理器 | [`context-manager.md`](agents\u002Forchestration\u002Fcontext-manager.md) | 压缩上下文，生成会话摘要 |\n| 工作流导演 | [`workflow-director.md`](agents\u002Forchestration\u002Fworkflow-director.md) | 多智能体流水线编排 |\n| 智能体安装器 | [`agent-installer.md`](agents\u002Forchestration\u002Fagent-installer.md) | 安装并配置智能体集合 |\n| 知识综合者 | [`knowledge-synthesizer.md`](agents\u002Forchestration\u002Fknowledge-synthesizer.md) | 压缩信息，构建知识图谱 |\n| 性能监控器 | [`performance-monitor.md`](agents\u002Forchestration\u002Fperformance-monitor.md) | 跟踪令牌使用情况，评估响应质量 |\n| 错误协调员 | [`error-coordinator.md`](agents\u002Forchestration\u002Ferror-coordinator.md) | 处理多智能体工作流中的错误 |\n| 多智能体协调员 | [`multi-agent-coordinator.md`](agents\u002Forchestration\u002Fmulti-agent-coordinator.md) | 并行执行智能体任务，合并输出 |\n\n### 研究与分析（11个智能体）\n\n| 智能体 | 文件 | 用途 |\n|-------|------|---------|\n| 研究分析师 | [`research-analyst.md`](agents\u002Fresearch-analysis\u002Fresearch-analyst.md) | 技术研究，证据综合 |\n| 竞争力分析师 | [`competitive-analyst.md`](agents\u002Fresearch-analysis\u002Fcompetitive-analyst.md) | 市场定位，功能对比 |\n| 趋势分析师 | [`trend-analyst.md`](agents\u002Fresearch-analysis\u002Ftrend-analyst.md) | 技术趋势预测 |\n| 数据研究员 | [`data-researcher.md`](agents\u002Fresearch-analysis\u002Fdata-researcher.md) | 数据分析，模式识别 |\n| 搜索专家 | [`search-specialist.md`](agents\u002Fresearch-analysis\u002Fsearch-specialist.md) | 信息检索，来源评估 |\n| 专利分析师 | [`patent-analyst.md`](agents\u002Fresearch-analysis\u002Fpatent-analyst.md) | 专利检索，现有技术分析，知识产权格局 |\n| 学术研究员 | [`academic-researcher.md`](agents\u002Fresearch-analysis\u002Facademic-researcher.md) | 文献综述，引文分析，方法论研究 |\n| 市场研究员 | [`market-researcher.md`](agents\u002Fresearch-analysis\u002Fmarket-researcher.md) | 市场规模测算，TAM\u002FSAM\u002FSOM，竞争情报 |\n| 安全研究员 | [`security-researcher.md`](agents\u002Fresearch-analysis\u002Fsecurity-researcher.md) | CVE分析，威胁建模，攻击面评估 |\n| 基准测试专家 | [`benchmarking-specialist.md`](agents\u002Fresearch-analysis\u002Fbenchmarking-specialist.md) | 性能基准测试，比较评估 |\n| 技术侦察员 | [`technology-scout.md`](agents\u002Fresearch-analysis\u002Ftechnology-scout.md) | 新兴技术评估，自建 vs 采购分析 |\n| agntk | [npx agntk](https:\u002F\u002Fgithub.com\u002FPhoenixrr2113\u002Fagntk) | 零配置AI智能体命令行工具，内置20余种工具，具备持久化记忆、Ollama自动检测功能，并提供免费套餐 |\n\n### 使用智能体\n\n在你的 `CLAUDE.md` 中引用某个智能体：\n\n```markdown\n## 智能体\n- 使用 `agents\u002Fcore-development\u002Ffullstack-engineer.md` 进行功能开发\n- 使用 `agents\u002Fquality-assurance\u002Fcode-reviewer.md` 进行代码评审\n- 使用 `agents\u002Fdata-ai\u002Fprompt-engineer.md` 进行提示词优化\n```\n\n---\n\n## 技能\n\n此仓库包含三十五个精选技能模块，并可通过[SkillKit市场](https:\u002F\u002Fagenstskills.com)访问**400,000+ 项额外技能**。每个技能模块都通过代码示例、反模式和检查清单，教授Claude Code领域的特定模式。\n\n| 技能 | 目录 | 教学内容 |\n|-------|-----------|-----------------|\n| TDD精通 | `skills\u002Ftdd-mastery\u002F` | 红‑绿‑重构、测试先行设计、覆盖率目标 |\n| API设计模式 | `skills\u002Fapi-design-patterns\u002F` | RESTful规范、版本控制、分页、错误响应 |\n| 数据库优化 | `skills\u002Fdatabase-optimization\u002F` | 查询计划、索引、N+1问题预防、连接池 |\n| 前端卓越 | `skills\u002Ffrontend-excellence\u002F` | 组件架构、状态管理、性能预算 |\n| 安全加固 | `skills\u002Fsecurity-hardening\u002F` | 输入验证、认证模式、密钥管理、CSP |\n| [Prism Scanner](https:\u002F\u002Fgithub.com\u002Faidongise-cell\u002Fprism-scanner) | `pip install prism-scanner` | 代理技能\u002F插件\u002FMCP安全扫描工具 — 39+条规则、AST污点跟踪、A-F等级评定 |\n| DevOps自动化 | `skills\u002Fdevops-automation\u002F` | 基础设施即代码、GitOps、监控、事件响应 |\n| 持续学习 | `skills\u002Fcontinuous-learning\u002F` | 会话摘要、学习日志、模式提取 |\n| React模式 | `skills\u002Freact-patterns\u002F` | Hooks、服务器组件、Suspense、错误边界 |\n| Python最佳实践 | `skills\u002Fpython-best-practices\u002F` | 类型提示、数据类、async\u002Fawait、打包 |\n| Go惯用法 | `skills\u002Fgolang-idioms\u002F` | 错误处理、接口、并发、项目布局 |\n| Django模式 | `skills\u002Fdjango-patterns\u002F` | DRF、ORM优化、信号、中间件 |\n| Spring Boot模式 | `skills\u002Fspringboot-patterns\u002F` | JPA、REST控制器、分层架构 |\n| Next.js精通 | `skills\u002Fnextjs-mastery\u002F` | App Router、RSC、ISR、服务器操作、中间件 |\n| GraphQL设计 | `skills\u002Fgraphql-design\u002F` | 架构设计、DataLoader、订阅、分页 |\n| Kubernetes运维 | `skills\u002Fkubernetes-operations\u002F` | 部署、Helm图表、HPA、故障排除 |\n| Docker最佳实践 | `skills\u002Fdocker-best-practices\u002F` | 多阶段构建、Compose、镜像优化 |\n| AWS云模式 | `skills\u002Faws-cloud-patterns\u002F` | Lambda、DynamoDB、CDK、S3事件处理 |\n| CI\u002FCD流水线 | `skills\u002Fci-cd-pipelines\u002F` | GitHub Actions、GitLab CI、矩阵构建 |\n| 微服务设计 | `skills\u002Fmicroservices-design\u002F` | 事件驱动架构、Saga模式、服务网格 |\n| TypeScript进阶 | `skills\u002Ftypescript-advanced\u002F` | 泛型、条件类型、映射类型、联合类型 |\n| Rust系统编程 | `skills\u002Frust-systems\u002F` | 所有权、trait、异步模式、错误处理 |\n| 提示工程 | `skills\u002Fprompt-engineering\u002F` | 思维链、少样本、结构化输出 |\n| MCP开发 | `skills\u002Fmcp-development\u002F` | MCP服务器工具、资源、传输设置 |\n| PostgreSQL优化 | `skills\u002Fpostgres-optimization\u002F` | EXPLAIN ANALYZE、索引、分区、JSONB |\n| Redis模式 | `skills\u002Fredis-patterns\u002F` | 缓存、限流、发布\u002F订阅、Streams、Lua脚本 |\n| 监控与可观测性 | `skills\u002Fmonitoring-observability\u002F` | OpenTelemetry、Prometheus、结构化日志 |\n| 认证模式 | `skills\u002Fauthentication-patterns\u002F` | JWT、OAuth2 PKCE、RBAC、会话管理 |\n| WebSocket与实时 | `skills\u002Fwebsocket-realtime\u002F` | Socket.io、SSE、重连机制、扩展性 |\n| 测试策略 | `skills\u002Ftesting-strategies\u002F` | 合同测试、快照测试、基于属性的测试 |\n| Git进阶 | `skills\u002Fgit-advanced\u002F` | 工作树、二分查找、交互式变基、钩子 |\n| 无障碍（WCAG） | `skills\u002Faccessibility-wcag\u002F` | ARIA模式、键盘导航、色彩对比度 |\n| 性能优化 | `skills\u002Fperformance-optimization\u002F` | 代码分割、图片优化、Core Web Vitals |\n| 移动开发 | `skills\u002Fmobile-development\u002F` | React Native、Flutter、响应式布局 |\n| 数据工程 | `skills\u002Fdata-engineering\u002F` | ETL管道、Spark、星型模式、数据质量 |\n| LLM集成 | `skills\u002Fllm-integration\u002F` | 流式传输、函数调用、RAG、成本优化 |\n\n### 社区技能\n\n| 技能 | 安装命令 | 教什么 |\n|-------|---------|------------------|\n| [Reepl - LinkedIn内容创作](https:\u002F\u002Fgithub.com\u002Freepl-io\u002Fskills) | `npx skillkit@latest install reepl-io\u002Fskills` | 18种用于LinkedIn内容管理的工具：草稿、发布、排期、语音简介、联系人、收藏、模板以及AI图像生成 |\n| [avoid-ai-writing](https:\u002F\u002Fgithub.com\u002Fconorbronsdon\u002Favoid-ai-writing) | `git clone https:\u002F\u002Fgithub.com\u002Fconorbronsdon\u002Favoid-ai-writing ~\u002F.claude\u002Fskills\u002Favoid-ai-writing` | AI写作模式检测与改写（21类，43种替换） |\n| [CCM](https:\u002F\u002Fgithub.com\u002Fdr5hn\u002Fccm) | `npx skills add dr5hn\u002Fccm@ccm` | 多账号管理、会话清理、健康检查、环境快照、Claude Code权限审计 |\n| [cc-inspect](https:\u002F\u002Fgithub.com\u002Fhowardpen9\u002Fcc-inspect) | `git clone https:\u002F\u002Fgithub.com\u002Fhowardpen9\u002Fcc-inspect ~\u002F.claude\u002Fskills\u002Fcc-inspect` | 在浏览器仪表盘中检查所有已安装的Claude Code技能、插件、MCP服务器、命令和钩子。具有作用域感知（用户\u002F项目\u002F本地），零依赖（纯bash + python3），生成自包含HTML |\n| [Edison](https:\u002F\u002Fgithub.com\u002Fkilnside\u002Fedison) | `git clone https:\u002F\u002Fgithub.com\u002Fkilnside\u002Fedison ~\u002F.claude\u002Fskills\u002Fedison` | 设计决策技能——介于头脑风暴与构建之间的阶段。以研究为先的渐进式深入探索，并配有可自动执行的规格说明。三种模式：检查、探索、审计 |\n| [claude-handoff](https:\u002F\u002Fgithub.com\u002FREMvisual\u002Fclaude-handoff) | `git clone https:\u002F\u002Fgithub.com\u002FREMvisual\u002Fclaude-handoff ~\u002F.claude\u002Fskills\u002Fhandoff` | 跨会话交接系统，具备链式追踪、上下文摘要以及Claude Code会话间结构化工作连续性功能 |\n| [ClawSec](https:\u002F\u002Fclawsec.cc) | 访问[clawsec.cc](https:\u002F\u002Fclawsec.cc) | 面向AI代理技能的安全审计平台——自动化五级评估、漏洞检测及配置审计，覆盖超过33,000个技能 |\n| [ClawSearch](https:\u002F\u002Fclawsearch.cc) | `npm install -g clawsearch` | 以安全为核心的技能发现引擎，提供信任评分、知识图谱，并配备预安装防护程序（`clawsearch-guard`），可在安装前对技能进行审查 |\n| [MUSE](https:\u002F\u002Fgithub.com\u002Fmyths-labs\u002Fmuse) | `git clone` + `.\u002Fscripts\u002Finstall.sh --tool claude` | 纯Markdown记忆操作系统，包含48个技能，支持跨对话记忆，设有8种角色并实现权限隔离。兼容6种AI编码工具。 |\n| [PM Pilot](https:\u002F\u002Fgithub.com\u002Fmshadmanrahman\u002Fpm-pilot) | `git clone https:\u002F\u002Fgithub.com\u002Fmshadmanrahman\u002Fpm-pilot ~\u002F.claude\u002Fskills\u002Fpm-pilot` | 面向项目经理的Claude Code：25个技能，用于会议准备、PRD编写、利益相关者情报收集、用户故事映射、客户旅程图绘制以及产品发现等任务 |\n| [产品经理技能](https:\u002F\u002Fgithub.com\u002FDigidai\u002Fproduct-manager-skills) | `clawhub install product-manager-skills` | 高级PM代理，涵盖6个知识领域、12个模板和30余种框架，涉及产品发现、战略制定、交付、SaaS指标、PM职业辅导（从IC到CPO）以及AI产品设计 |\n| [OpenDivination](https:\u002F\u002Fgithub.com\u002Famenti-labs\u002Fopendivination) | `pipx install opendivination && npx skills add amenti-labs\u002Fopendivination --skill divination` | 塔罗牌和易经技能，具备可审计的熵源证明，支持通过计算机随机数生成器、QRNG API和本地硬件等多种方式引导源设置，并可选共振模式 |\n| [pumpclaw](https:\u002F\u002Fgithub.com\u002Fchainstacklabs\u002Fpumpclaw) | `git clone https:\u002F\u002Fgithub.com\u002Fchainstacklabs\u002Fpumpclaw ~\u002F.claude\u002Fskills\u002Fpumpclaw` | 面向Solana上pump.fun代币交易的代理技能——买入、卖出、发行代币、钱包管理、模拟运行、滑点检查以及通过pumpfun-cli进行PumpSwap AMM迁移 |\n| [x-twitter-scraper](https:\u002F\u002Fgithub.com\u002FXquik-dev\u002Fx-twitter-scraper) | `npx skills add Xquik-dev\u002Fx-twitter-scraper` | 面向AI编码代理的X API和Twitter抓取技能——推文搜索、用户查找、粉丝提取、互动指标、抽奖活动、热门话题、账户监控以及19种数据提取工具 |\n| [claude-code-marketing-skills](https:\u002F\u002Fgithub.com\u002Fcognyai\u002Fclaude-code-marketing-skills) | `git clone https:\u002F\u002Fgithub.com\u002Fcognyai\u002Fclaude-code-marketing-skills ~\u002F.claude\u002Fskills\u002Fclaude-code-marketing-skills` | AI驱动的营销技能——SEO审计、着陆页评审、竞争对手分析、广告文案撰写、潜在客户资格评估。免费，无需注册账号。兼容Claude Code、Cursor、Windsurf |\n| [claude-skills](https:\u002F\u002Fgithub.com\u002Falirezarezvani\u002Fclaude-skills) | `git clone` | 涵盖9大领域的192个生产就绪技能（工程、营销、产品、合规、高层咨询），配备254个Python自动化工具。获5,300+星 |\n| [faf-skills](https:\u002F\u002Fgithub.com\u002FWolfe-Jam\u002Ffaf-skills) | `git clone` | 31个Claude Code技能，用于持久化项目上下文（.faf, `application\u002Fvnd.faf+yaml`）。包括评分、同步、测试、发布、MCP服务器创建、架构文档等。Anthropic MCP #2759。 |\n| [floom](https:\u002F\u002Fgithub.com\u002Ffloomhq\u002Ffloom) | `git clone https:\u002F\u002Fgithub.com\u002Ffloomhq\u002Ffloom.git ~\u002F.claude\u002Fskills\u002Ffloom-repo && ~\u002F.claude\u002Fskills\u002Ffloom-repo\u002Fscripts\u002Fsetup` | 将Python脚本部署为云端自动化。`\u002Ffloom`会适配代码，在沙盒中测试，并以自动生成的Web界面、REST API和MCP端点进行部署 |\n| [n8n-skills](https:\u002F\u002Fgithub.com\u002Fczlonkowski\u002Fn8n-skills) | `git clone` | 7个互补技能，用于构建生产就绪的n8n工作流。覆盖525+节点、2,653+模板。获3,400+星 |\n| [nv:context](https:\u002F\u002Fgithub.com\u002Fjohnnichev\u002Fnv-context) | `npx skills add johnnichev\u002Fnv-context -g -y` | 最先进的AI代理上下文工程。自动发现工具，审计代码中的隐患，对6层杠杆因素（仓库结构、工具集成、内存、钩子、会话、角色设定）进行评分，并生成生产就绪的CLAUDE.md\u002FAGENTS.md配置文件 |\n| [claude-code-best-practice](https:\u002F\u002Fgithub.com\u002Fshanraisshan\u002Fclaude-code-best-practice) | `git clone` | Claude Code配置的全面参考实现——技能、子代理、钩子、命令，并附有实用示例。获17,400+星 |\n| [ADHX](https:\u002F\u002Fgithub.com\u002Fitsmemeworks\u002Fadhx) | `\u002Fplugin marketplace add itsmemeworks\u002Fadhx` 或 `curl -sL https:\u002F\u002Fraw.githubusercontent.com\u002Fitsmemeworks\u002Fadhx\u002Fmain\u002Fskills\u002Fadhx\u002FSKILL.md -o ~\u002F.claude\u002Fskills\u002Fadhx\u002FSKILL.md` | 将任何X\u002FTwitter帖子提取为干净的LLM友好型JSON——无需爬虫，适用于推文和完整的X文章 |\n| [SkillNav](https:\u002F\u002Fgithub.com\u002Fskillnav-dev\u002Fskillnav-skill) | `clawhub install HuiW86\u002Fskillnav` | 搜索3,900+个MCP服务器及其安装命令，获取每日AI简报，查询arXiv论文，并发现热门工具——全部以中文呈现。由skillnav.dev编辑团队精选 |\n| [moyu](https:\u002F\u002Fgithub.com\u002Fuucz\u002Fmoyu) | `claude skill install --url https:\u002F\u002Fgithub.com\u002Fuucz\u002Fmoyu --skill moyu` | 反过度工程技能，教导AI保持克制。5种变体（标准\u002F轻量\u002F严格\u002F英\u002F日），适用于10个平台。基准测试显示，相比基线可减少66%的代码量 |\n| [naming](https:\u002F\u002Fgithub.com\u002Fglacierphonk\u002Fnaming) | `git clone https:\u002F\u002Fgithub.com\u002Fglacierphonk\u002Fnaming ~\u002F.claude\u002Fskills\u002Fnaming` | 以隐喻为导向的产品、SaaS、品牌、机器人和开源项目命名。生成令人难忘且富有意义的名称 |\n| [Obsidian主题设计师](https:\u002F\u002Fgithub.com\u002FXiangyuSu611\u002Fobsidian-theme-designer) | `git clone https:\u002F\u002Fgithub.com\u002FXiangyuSu611\u002Fobsidian-theme-designer ~\u002F.claude\u002Fskills\u002Fobsidian-theme-designer` | 在浏览器中可视化设计Obsidian主题——风格方向、色彩搭配、字体展示、双色浅\u002F深模式预览，以及自动字体安装。无需CSS知识 |\n| [Claudify](https:\u002F\u002Fclaudify.tech) | `npx create-claudify` | 为Claude Code打造的完整操作系统，包含31个类别下的1,727个技能，9个具有持久记忆的专业代理，21条斜杠命令，以及自动化质量检查 |\n| [Axiom](https:\u002F\u002Fgithub.com\u002FGuipetris\u002Faxiom) | `npx skills add Guipetris\u002Faxiom` | 合同约束的自主流水线：从创意到PR共8个阶段，遵循EARS要求、ATDD（先测试后实现）、RALPLAN-DR审议与ADR记录、3名独立平行评审员、分类自我修正（语法\u002F逻辑\u002F架构\u002F模糊）、跨运行项目记忆以及阶段回滚。除git和gh外无其他依赖 |\n| [AuraKit](https:\u002F\u002Fgithub.com\u002Fsmorky850612\u002FAurakit) | `npx @smorky85\u002Faurakit` | 全栈开发技能，包含37种模式（构建\u002F修复\u002F清洁\u002F部署\u002F评审+32种扩展），实施6层OWASP安全防护，配备30个运行时钩子、10名特定语言的评审员，以及Instinct学习引擎。跨平台：Claude Code、Codex CLI、Cursor、Windsurf、Manus。 |\n| [StitchFlow](https:\u002F\u002Fgithub.com\u002Fyshishenya\u002Fstitchflow) | `git clone https:\u002F\u002Fgithub.com\u002Fyshishenya\u002Fstitchflow && cd stitchflow && bash install.sh --target all` | 跨代理的Stitch UI设计套件，可将需求文档和原型转化为界面、变体、Tailwind友好的HTML以及截图 |\n\n| [成本优化器](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-cost-optimizer) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-cost-optimizer` | 通过智能模型路由、上下文压缩、心跳调优、使用情况报告和配置生成，节省60–80%的AI token成本 |\n| [夜间工作者](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-overnight-worker) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-overnight-worker` | 自主夜间工作代理——睡前分配任务，借助智能任务分解、网络研究和推送通知，早晨即可获得结构化结果 |\n| [智能PR评审](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-smart-pr-review) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-smart-pr-review` | 带有6层深度评审、反方模式以及针对TS\u002FJS、Python、Go、Rust的标准输出格式（MUST FIX \u002F SHOULD FIX \u002F SUGGESTION）的意见型AI代码评审工具 |\n| [DevOps代理](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-devops-agent) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-devops-agent` | 一键部署、监控设置（Prometheus+Grafana）、定时备份和故障诊断，采用安全优先设计，包含确认提示、试运行和快照回滚功能 |\n| [中文内容矩阵](https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-cn-content-matrix) | `git clone https:\u002F\u002Fgithub.com\u002Ffullstackcrew-alpha\u002Fskill-cn-content-matrix` | 面向小红书、微信、抖音和B站的中文多平台内容生成器，具备真实风格迁移、合规性审查和敏感词检测功能 |\n| [D3.js可视化](https:\u002F\u002Fgithub.com\u002Fchrisvoncsefalvay\u002Fclaude-d3js-skill) | `git clone` | 从Claude Code生成交互式D3.js数据可视化 |\n| [Google Drive – Memyard](https:\u002F\u002Fgithub.com\u002Fzagmoai\u002Fpublic-google-drive) | `git clone https:\u002F\u002Fgithub.com\u002Fzagmoai\u002Fpublic-google-drive ~\u002F.claude\u002Fskills\u002Fpublic-google-drive` | 无需登录即可创建和编辑Google Docs与Sheets。文档托管在Memyard上，并提供可分享链接。无需API密钥或OAuth授权 |\n| [iOS模拟器](https:\u002F\u002Fgithub.com\u002Fconorluddy\u002Fios-simulator-skill) | `git clone` | 与iOS模拟器交互，用于移动设备测试和截图捕获 |\n| [Playwright自动化](https:\u002F\u002Fgithub.com\u002Flackeyjb\u002Fplaywright-skill) | `git clone` | 用于端到端测试和网页交互的浏览器自动化技能 |\n| [基于Sigma规则的威胁狩猎](https:\u002F\u002Fgithub.com\u002Fjthack\u002Fthreat-hunting-with-sigma-rules-skill) | `git clone` | 基于Sigma规则的威胁检测与安全日志分析 |\n| [FFUF Web模糊测试](https:\u002F\u002Fgithub.com\u002Fjthack\u002Fffuf_claude_skill) | `git clone` | 通过FFUF进行Web漏洞模糊测试和安全评估 |\n| [Markdown转EPUB](https:\u002F\u002Fgithub.com\u002Fsmerchek\u002Fclaude-epub-skill) | `git clone` | 将Markdown文件转换为EPUB电子书 |\n| [CSV数据摘要器](https:\u002F\u002Fgithub.com\u002Fcoffeefuelbump\u002Fcsv-data-summarizer-claude-skill) | `git clone` | 从CSV数据文件中生成洞察和摘要 |\n| [tapestry](https:\u002F\u002Fgithub.com\u002Fmichalparkola\u002Ftapestry-skills-for-claude-code) | `git clone` | 知识网络、迭代学习、文章提取和YouTube字幕处理 |\n| [j4rk0r\u002Fclaude-skills](https:\u002F\u002Fgithub.com\u002Fj4rk0r\u002Fclaude-skills) | `npx skills add j4rk0r\u002Fclaude-skills --yes --global` | 3项专家级技能：skill-guard（9层安全审计员）、skill-advisor（带差距分析的智能路由）和skill-learner（持续错误修正）。所有技能在skill-judge上均获得A+ 120\u002F120分 |\n| [SciAgent-Skills](https:\u002F\u002Fgithub.com\u002Fjaechang-hits\u002FSciAgent-Skills) | `git clone https:\u002F\u002Fgithub.com\u002Fjaechang-hits\u002FSciAgent-Skills` | 面向Claude Code的197项生命科学技能，涵盖基因组学、蛋白质组学、药物研发等领域。BixBench准确率为92.0%。 |\n| [manage-skills](https:\u002F\u002Fgithub.com\u002Fumutbozdag\u002Fagent-skills-manager\u002Ftree\u002Fmain\u002Fskills\u002Fmanage-skills) | `cp -r skills\u002Fmanage-skills ~\u002F.claude\u002Fskills\u002Fmanage-skills` | 在11种工具（Cursor、Claude、Windsurf、Copilot、Codex、Cline、Aider、Continue、Roo Code、Augment）中，通过终端发现、创建、编辑、切换、复制、移动和删除代理技能 |\n\n\n\n### 安装技能\n\n**通过SkillKit浏览并安装**（推荐）：\n\n```bash\nnpx skillkit@latest install claude-code-toolkit\u002Ftdd-mastery\n```\n\n### SkillKit市场提供超过15,000项技能\n\n该工具包包含35项精选技能。如需访问涵盖各领域的**40万多项额外技能**，请使用[SkillKit](https:\u002F\u002Fagenstskills.com)：\n\n```bash\nnpx skillkit@latest                    # 启动交互式TUI\nnpx skillkit@latest search \"react\"     # 搜索40万多项技能\nnpx skillkit@latest recommend          # AI驱动的技能推荐\n```\n\n完整市场可在[agenstskills.com](https:\u002F\u002Fagenstskills.com)浏览。SkillKit支持包括Claude Code、Cursor、Codex、Gemini CLI等在内的32余款AI编码代理。\n\n---\n\n## 命令\n\n四十二个斜杠命令，分为八类。将这些命令放入项目的`.claude\u002Fcommands\u002F`目录中。\n\n### Git（7条命令）\n\n| 命令 | 文件 | 描述 |\n|---------|------|-------------|\n| `\u002Fcommit` | [`commit.md`](commands\u002Fgit\u002Fcommit.md) | 根据暂存更改生成规范化的提交信息 |\n| `\u002Fpr-create` | [`pr-create.md`](commands\u002Fgit\u002Fpr-create.md) | 创建带有摘要、测试计划和标签的拉取请求 |\n| `\u002Fchangelog` | [`changelog.md`](commands\u002Fgit\u002Fchangelog.md) | 根据提交历史生成变更日志 |\n| `\u002Frelease` | [`release.md`](commands\u002Fgit\u002Frelease.md) | 创建带有自动生成备注的标记版本 |\n| `\u002Fworktree` | [`worktree.md`](commands\u002Fgit\u002Fworktree.md) | 设置Git工作树以支持并行开发 |\n| `\u002Ffix-issue` | [`fix-issue.md`](commands\u002Fgit\u002Ffix-issue.md) | 根据编号修复GitHub问题 |\n| `\u002Fpr-review` | [`pr-review.md`](commands\u002Fgit\u002Fpr-review.md) | 对拉取请求进行结构化反馈评审 |\n\n### 测试（6条命令）\n\n| 命令 | 文件 | 描述 |\n|---------|------|-------------|\n| `\u002Ftdd` | [`tdd.md`](commands\u002Ftesting\u002Ftdd.md) | 测试驱动开发流程 |\n| `\u002Ftest-coverage` | [`test-coverage.md`](commands\u002Ftesting\u002Ftest-coverage.md) | 分析覆盖率并建议缺失的测试用例 |\n| `\u002Fe2e` | [`e2e.md`](commands\u002Ftesting\u002Fe2e.md) | 生成端到端测试场景 |\n| `\u002Fintegration-test` | [`integration-test.md`](commands\u002Ftesting\u002Fintegration-test.md) | 为API端点生成集成测试 |\n| `\u002Fsnapshot-test` | [`snapshot-test.md`](commands\u002Ftesting\u002Fsnapshot-test.md) | 生成快照\u002F黄金文件测试 |\n| `\u002Ftest-fix` | [`test-fix.md`](commands\u002Ftesting\u002Ftest-fix.md) | 诊断并修复失败的测试 |\n\n### 架构（6 个命令）\n\n| 命令 | 文件 | 描述 |\n|---------|------|-------------|\n| `\u002Fplan` | [`plan.md`](commands\u002Farchitecture\u002Fplan.md) | 创建包含风险评估的实施计划 |\n| `\u002Frefactor` | [`refactor.md`](commands\u002Farchitecture\u002Frefactor.md) | 结构化的代码重构流程 |\n| `\u002Fmigrate` | [`migrate.md`](commands\u002Farchitecture\u002Fmigrate.md) | 框架或库的迁移 |\n| `\u002Fadr` | [`adr.md`](commands\u002Farchitecture\u002Fadr.md) | 编写架构决策记录 |\n| `\u002Fdiagram` | [`diagram.md`](commands\u002Farchitecture\u002Fdiagram.md) | 从代码生成 Mermaid 图表 |\n| `\u002Fdesign-review` | [`design-review.md`](commands\u002Farchitecture\u002Fdesign-review.md) | 进行结构化的设计评审 |\n\n### 文档（5 个命令）\n\n| 命令 | 文件 | 描述 |\n|---------|------|-------------|\n| `\u002Fdoc-gen` | [`doc-gen.md`](commands\u002Fdocumentation\u002Fdoc-gen.md) | 从代码生成文档 |\n| `\u002Fupdate-codemap` | [`update-codemap.md`](commands\u002Fdocumentation\u002Fupdate-codemap.md) | 更新项目代码地图 |\n| `\u002Fapi-docs` | [`api-docs.md`](commands\u002Fdocumentation\u002Fapi-docs.md) | 从路由处理器生成 API 文档 |\n| `\u002Fonboard` | [`onboard.md`](commands\u002Fdocumentation\u002Fonboard.md) | 为新开发者创建入职指南 |\n| `\u002Fmemory-bank` | [`memory-bank.md`](commands\u002Fdocumentation\u002Fmemory-bank.md) | 更新 CLAUDE.md 记忆库 |\n\n### 安全（5 个命令）\n\n| 命令 | 文件 | 描述 |\n|---------|------|-------------|\n| `\u002Faudit` | [`audit.md`](commands\u002Fsecurity\u002Faudit.md) | 对代码和依赖项进行安全审计 |\n| `\u002Fhardening` | [`hardening.md`](commands\u002Fsecurity\u002Fhardening.md) | 应用安全加固措施 |\n| `\u002Fsecrets-scan` | [`secrets-scan.md`](commands\u002Fsecurity\u002Fsecrets-scan.md) | 扫描泄露的密钥和凭据 |\n| `\u002Fcsp` | [`csp.md`](commands\u002Fsecurity\u002Fcsp.md) | 生成内容安全策略头 |\n| `\u002Fdependency-audit` | [`dependency-audit.md`](commands\u002Fsecurity\u002Fdependency-audit.md) | 审计依赖项中的漏洞 |\n\n### 重构（5 个命令）\n\n| 命令 | 文件 | 描述 |\n|---------|------|-------------|\n| `\u002Fdead-code` | [`dead-code.md`](commands\u002Frefactoring\u002Fdead-code.md) | 查找并移除无用代码 |\n| `\u002Fsimplify` | [`simplify.md`](commands\u002Frefactoring\u002Fsimplify.md) | 降低当前文件的复杂度 |\n| `\u002Fextract` | [`extract.md`](commands\u002Frefactoring\u002Fextract.md) | 提取函数、组件或模块 |\n| `\u002Frename` | [`rename.md`](commands\u002Frefactoring\u002Frename.md) | 在整个代码库中重命名符号 |\n| `\u002Fcleanup` | [`cleanup.md`](commands\u002Frefactoring\u002Fcleanup.md) | 移除无用代码和未使用的导入 |\n\n### DevOps（5 个命令）\n\n| 命令 | 文件 | 描述 |\n|---------|------|-------------|\n| `\u002Fdockerfile` | [`dockerfile.md`](commands\u002Fdevops\u002Fdockerfile.md) | 生成优化的 Dockerfile |\n| `\u002Fci-pipeline` | [`ci-pipeline.md`](commands\u002Fdevops\u002Fci-pipeline.md) | 生成 CI\u002FCD 管道配置 |\n| `\u002Fk8s-manifest` | [`k8s-manifest.md`](commands\u002Fdevops\u002Fk8s-manifest.md) | 生成 Kubernetes 清单文件 |\n| `\u002Fdeploy` | [`deploy.md`](commands\u002Fdevops\u002Fdeploy.md) | 部署到配置的环境 |\n| `\u002Fmonitor` | [`monitor.md`](commands\u002Fdevops\u002Fmonitor.md) | 设置监控和告警 |\n\n### 工作流（3 个命令）\n\n| 命令 | 文件 | 描述 |\n|---------|------|-------------|\n| `\u002Fcheckpoint` | [`checkpoint.md`](commands\u002Fworkflow\u002Fcheckpoint.md) | 保存会话进度和上下文 |\n| `\u002Fwrap-up` | [`wrap-up.md`](commands\u002Fworkflow\u002Fwrap-up.md) | 以总结和经验结束会话 |\n| `\u002Forchestrate` | [`orchestrate.md`](commands\u002Fworkflow\u002Forchestrate.md) | 运行多智能体工作流管道 |\n\n### 使用命令\n\n复制到您的项目：\n\n```bash\ncp -r commands\u002F .claude\u002Fcommands\u002F\n```\n\n然后在 Claude Code 中调用：\n\n```\n\u002Fcommit\n\u002Ftdd src\u002Futils\u002Fparser.ts\n\u002Faudit\n\u002Forchestrate feature \"Add user authentication\"\n```\n\n---\n\n## 钩子\n\n二十个钩子脚本，覆盖所有八个 Claude Code 生命周期事件。将 `hooks.json` 放入您的 `.claude\u002F` 目录。\n\n### 钩子脚本\n\n| 脚本 | 触发条件 | 目的 |\n|--------|---------|---------|\n| `session-start.js` | 会话开始 | 加载项目上下文，检测包管理器 |\n| `session-end.js` | 会话结束 | 保存会话状态以便下次使用 |\n| `context-loader.js` | 会话开始 | 加载 CLAUDE.md、git 状态、待办事项 |\n| `learning-log.js` | 会话结束 | 提取并保存会话学习成果 |\n| `pre-compact.js` | 预压缩 | 在压缩前保存重要上下文 |\n| [`smart-approve.py`](https:\u002F\u002Fgithub.com\u002Fliberzon\u002Fclaude-hooks) | 工具使用前（Bash） | 将复合 Bash 命令（&&、||、;、|、$()）分解为子命令，并逐一检查是否符合允许或禁止模式 |\n| `block-dev-server.js` | 工具使用前（Bash） | 阻止在 tmux 外部运行开发服务器命令 |\n| `pre-push-check.js` | 工具使用前（Bash） | 推送前验证分支和远程仓库 |\n| `block-md-creation.js` | 工具使用前（写入） | 阻止不必要的 .md 文件创建 |\n| `commit-guard.js` | 工具使用前（Bash） | 验证规范化的提交信息 |\n| `secret-scanner.js` | 工具使用前（写入\u002F编辑） | 阻止包含密钥的文件 |\n| `post-edit-check.js` | 工具使用后（写入\u002F编辑） | 文件编辑后运行 linter |\n| `auto-test.js` | 工具使用后（写入\u002F编辑） | 编辑后运行相关测试 |\n| `type-check.js` | 工具使用后（写入\u002F编辑） | 编辑后进行 TypeScript 类型检查 |\n| `lint-fix.js` | 工具使用后（写入\u002F编辑） | 自动修复 lint 问题 |\n| `bundle-check.js` | 工具使用后（Bash） | 构建后检查打包大小 |\n| `suggest-compact.js` | 工具使用后（Bash） | 在编辑间隔时建议压缩 |\n| `stop-check.js` | 停止 | 如果代码被修改，提醒运行测试 |\n| `notification-log.js` | 通知 | 记录通知以便后续查看 |\n| `prompt-check.js` | 用户提交提示 | 检测模糊提示，建议澄清 |\n\n### 相关 SDK\n\n如果您更倾向于使用类型化、可通过 npm 安装的基础框架来编写钩子，而不是原始脚本：\n\n- [claude-code-hooks](https:\u002F\u002Fgithub.com\u002FPayshak\u002Fclaude-code-hooks) — TypeScript SDK，包含 `defineHook()`、所有 5 个钩子事件的类型化事件负载、响应构建器以及可单元测试的 `.handle()` 方法。无依赖项。\n\n### 安装钩子\n\n```bash\ncp hooks\u002Fhooks.json .claude\u002Fhooks.json\ncp -r hooks\u002Fscripts\u002F .claude\u002Fhooks\u002Fscripts\u002F\n```\n\n### 使用 cc-safe-setup 快速设置\n\n只需一条命令即可安装 8 个经过生产验证的安全钩子——破坏性命令阻止器、分支推送保护、防止密钥泄露、语法验证等：\n\n```bash\nnpx cc-safe-setup\n```\n\n包括 `--audit`（为您的设置打分 0-100）、`--scan`（检测技术栈，推荐钩子）和 `--verify`（测试每个钩子）。详情请参阅 [cc-safe-setup](https:\u002F\u002Fgithub.com\u002Fyurukusa\u002Fcc-safe-setup)。\n\n---\n\n## 规则\n\n十五条编码规则，用于强制执行一致的代码风格。可将其添加到 `.claude\u002Frules\u002F` 目录下，或在 `CLAUDE.md` 中引用。\n\n| 规则 | 文件 | 强制内容 |\n|------|------|-----------------|\n| 编码风格 | [`coding-style.md`](rules\u002Fcoding-style.md) | 命名规范、文件组织、导入顺序 |\n| Git 工作流 | [`git-workflow.md`](rules\u002Fgit-workflow.md) | 分支策略、提交格式、拉取请求流程 |\n| 测试 | [`testing.md`](rules\u002Ftesting.md) | 测试结构、覆盖率目标、模拟指南 |\n| 安全 | [`security.md`](rules\u002Fsecurity.md) | 输入验证、敏感信息管理、参数化查询 |\n| 性能 | [`performance.md`](rules\u002Fperformance.md) | 懒加载、缓存、包优化 |\n| 文档 | [`documentation.md`](rules\u002Fdocumentation.md) | 公共 API 的 JSDoc 注释、内联注释规范 |\n| 错误处理 | [`error-handling.md`](rules\u002Ferror-handling.md) | 显式错误处理、类型化错误、禁止空 catch |\n| 代理 | [`agents.md`](rules\u002Fagents.md) | 代理设计模式、任务交接协议 |\n| API 设计 | [`api-design.md`](rules\u002Fapi-design.md) | REST 约定、状态码、版本控制 |\n| 可访问性 | [`accessibility.md`](rules\u002Faccessibility.md) | WCAG 2.2 标准、ARIA 属性、语义化 HTML |\n| 数据库 | [`database.md`](rules\u002Fdatabase.md) | 查询模式、数据库迁移、避免 N+1 查询问题 |\n| 依赖管理 | [`dependency-management.md`](rules\u002Fdependency-management.md) | 版本锁定、依赖审计、更新策略 |\n| 代码审查 | [`code-review.md`](rules\u002Fcode-review.md) | 审查检查清单、批准标准 |\n| 监控 | [`monitoring.md`](rules\u002Fmonitoring.md) | 日志记录标准、指标监控、告警机制 |\n| 命名 | [`naming.md`](rules\u002Fnaming.md) | 各语言的命名规范 |\n\n---\n\n## 模板\n\n- **[claude-code-blueprint](https:\u002F\u002Fgithub.com\u002Ffaizkhairi\u002Fclaude-code-blueprint)** - 经过实战检验的 Claude Code 高级用户参考架构。包含专用代理、模型分层、自然语言处理能力、生命周期钩子、路径作用域规则、初始预设、基准测试、实战案例以及跨工具映射。无依赖，MIT 许可证。\n- **[The CLAUDE.md Bible](https:\u002F\u002Fechochime3.gumroad.com\u002Fl\u002Fclaudemd-bible)** - 25 种针对不同技术栈的 CLAUDE.md 配置，涵盖 React、Next.js、FastAPI、Django、Svelte、Chrome 扩展、CLI 工具、MCP 服务器等。每种配置包含 80–150 行经过验证的规则，针对特定技术栈的常见模式和陷阱，并附有大师级指南。\n- **[TemplateClaw](https:\u002F\u002Fgithub.com\u002Fjeromwolf\u002Ftemplateclaw)** - 开发者模板中心及项目脚手架工具，打包为 Claude Code 插件。包含 6 大类共 32 个生产就绪模板（着陆页、仪表盘、UI 组件、开发方法论、项目初始化、重构）。提供 7 条 Slash 命令，包括 `\u002Ftemplateclaw`、`\u002Ftemplateclaw-landing`、`\u002Ftemplateclaw-dashboard`、`\u002Ftemplateclaw-ui` 等。兼容 Claude Code、Codex、Gemini CLI 和 Cursor。MIT 许可证。\n\n\n七种适用于不同项目类型的 CLAUDE.md 模板。\n\n| 模板 | 文件 | 使用场景 |\n|----------|------|----------|\n| 极简 | [`minimal.md`](templates\u002Fclaude-md\u002Fminimal.md) | 小型项目、脚本、快速原型 |\n| 标准 | [`standard.md`](templates\u002Fclaude-md\u002Fstandard.md) | 多数项目——涵盖偏好、规则和工作流 |\n| 全面 | [`comprehensive.md`](templates\u002Fclaude-md\u002Fcomprehensive.md) | 大型代码库，具有详细约定 |\n| 单体仓库 | [`monorepo.md`](templates\u002Fclaude-md\u002Fmonorepo.md) | Turborepo\u002FNx 单体仓库，包含多个包 |\n| 企业级 | [`enterprise.md`](templates\u002Fclaude-md\u002Fenterprise.md) | 大型团队，需满足合规性和 SSO 要求 |\n| Python 项目 | [`python-project.md`](templates\u002Fclaude-md\u002Fpython-project.md) | FastAPI\u002FDjango Python 项目 |\n| 全栈应用 | [`fullstack-app.md`](templates\u002Fclaude-md\u002Ffullstack-app.md) | Next.js + API 的全栈应用 |\n\n```bash\ncp templates\u002Fclaude-md\u002Fstandard.md CLAUDE.md\n```\n\n---\n\n## MCP 配置\n\n十三种精选的 Model Context Protocol 服务器配置。\n\n| 配置 | 文件 | 包含的服务器 |\n|--------|------|-----------------|\n| 推荐 | [`recommended.json`](mcp-configs\u002Frecommended.json) | 14 个通用开发必备服务器 |\n| 全栈 | [`fullstack.json`](mcp-configs\u002Ffullstack.json) | 文件系统、GitHub、Postgres、Redis、Puppeteer |\n| Kubernetes | [`kubernetes.json`](mcp-configs\u002Fkubernetes.json) | kubectl-mcp-server、Docker、GitHub |\n| 数据科学 | [`data-science.json`](mcp-configs\u002Fdata-science.json) | Jupyter、SQLite、PostgreSQL、文件系统 |\n| 前端 | [`frontend.json`](mcp-configs\u002Ffrontend.json) | Puppeteer、Figma、Storybook |\n| 加密货币 \u002F DeFi | [`crypto-defi.json`](mcp-configs\u002Fcrypto-defi.json) | defi-mcp、文件系统、Fetch、内存 |\n| DevOps | [`devops.json`](mcp-configs\u002Fdevops.json) | AWS、Docker、GitHub、Terraform、Sentry |\n| 研究 | [`research.json`](mcp-configs\u002Fresearch.json) | BGPT 科学论文、Brave Search、Fetch、内存、文件系统 |\n| 可观测性 | [`observability.json`](mcp-configs\u002Fobservability.json) | Iris 评估与可观ility 用于代理追踪、质量评估和成本监控 |\n| 安全 | [`security.json`](mcp-configs\u002Fsecurity.json) | Ghidra 逆向工程、Snyk 漏洞扫描 |\n| 设计 | [`design.json`](mcp-configs\u002Fdesign.json) | Figma 设计上下文、Blender 3D 自动化 |\n| 工作流自动化 | [`workflow-automation.json`](mcp-configs\u002Fworkflow-automation.json) | n8n 工作流构建器、Pipedream 集成 |\n| 移动 | [`mobile.json`](mcp-configs\u002Fmobile.json) | Android ADB 自动化、Xcode 构建工具 |\n| 金融 | [`finance.json`](mcp-configs\u002Ffinance.json) | Chart Library 模式识别、Fetch、内存 |\n\n---\n\n## 上下文\n\n五种上下文模式，用于根据不同任务配置 Claude Code 的行为。\n\n| 上下文 | 文件 | 重点 |\n|---------|------|-------|\n| 开发 | [`dev.md`](contexts\u002Fdev.md) | 快速迭代、遵循模式、边写代码边测试 |\n| 代码审查 | [`review.md`](contexts\u002Freview.md) | 检查逻辑、安全性、边界情况 |\n| 研究 | [`research.md`](contexts\u002Fresearch.md) | 评估工具、比较替代方案、记录发现 |\n| 调试 | [`debug.md`](contexts\u002Fdebug.md) | 复现问题、提出假设、修复根本原因、回归测试 |\n| 部署 | [`deploy.md`](contexts\u002Fdeploy.md) | 部署前检查清单、先部署到预发布环境、回滚标准 |\n\n---\n\n## 示例\n\n三个演示真实工具使用情况的示例。\n\n| 示例 | 文件 | 描述 |\n|---------|------|-------------|\n| 会话工作流 | [`session-workflow.md`](examples\u002Fsession-workflow.md) | 端到端高效开发会话 |\n| 多代理流水线 | [`multi-agent-pipeline.md`](examples\u002Fmulti-agent-pipeline.md) | 链式调用代理实现 Stripe 计费功能 |\n| 项目初始化 | [`project-setup.md`](examples\u002Fproject-setup.md) | 使用完整工具集设置新项目 |\n\n---\n\n## 安装\n\n```bash\nbash setup\u002Finstall.sh\n```\n\n交互式安装程序会克隆仓库、创建配置符号链接并安装插件。\n\n---\n\n## 项目结构\n\n```\nclaude-code-toolkit\u002F               850+ 个文件\n  plugins\u002F                         150+ 个插件（220 个命令文件 + 外部插件）\n  agents\u002F                          10 个类别中共 135 个智能体\n    core-development\u002F              13 个智能体\n    language-experts\u002F              25 个智能体\n    infrastructure\u002F                11 个智能体\n    quality-assurance\u002F             10 个智能体\n    data-ai\u002F                       15 个智能体\n    developer-experience\u002F          15 个智能体\n    specialized-domains\u002F           15 个智能体\n    business-product\u002F              12 个智能体\n    orchestration\u002F                 8 个智能体\n    research-analysis\u002F             11 个智能体\n  skills\u002F                          35 个 SKILL.md 文件\n  commands\u002F                        8 个类别中共 42 个命令\n  hooks\u002F\n    hooks.json                     25 条钩子配置\n    scripts\u002F                       19 个 Node.js 脚本\n  rules\u002F                           15 条编码规范\n  templates\u002Fclaude-md\u002F             7 个 CLAUDE.md 模板\n  mcp-configs\u002F                     8 种服务器配置\n  contexts\u002F                        5 种上下文模式\n  examples\u002F                        3 个操作指南示例\n  setup\u002F                           交互式安装程序\n```\n\n---\n\n## 配套应用与图形界面\n\n| 名称 | 星级 | 描述 |\n|------|-------|-------------|\n| [ctrl](https:\u002F\u002Fctrl.bulletproof.sh) | - | 一款像素风办公环境，实时可视化 AI 编码代理（Claude Code、Codex、Gemini）的工作状态。独立 Web 应用 + 守护进程，支持中继共享以实现远程访问，并且多代理兼容 |\n| [Opcode](https:\u002F\u002Fgithub.com\u002Fwinfunc\u002Fopcode) | 21,000+ | 基于 Tauri 2 的桌面 GUI 和工具包，专为 Claude Code 打造——可通过可视化编辑器创建自定义代理，提供使用情况分析及 MCP 集成 |\n| [CloudCLI](https:\u002F\u002Fgithub.com\u002Fsiteboon\u002Fclaudecodeui) | 8,400+ | 免费开源的 Web\u002F移动端 UI，适用于 Claude Code、Cursor CLI 和 Codex。响应式设计，集成终端、文件管理器和 Git 管理器 |\n| [Companion](https:\u002F\u002Fgithub.com\u002FThe-Vibe-Company\u002Fcompanion) | 2,200+ | Claude Code 和 Codex 的 Web 及移动 UI。可启动并行会话、流式传输响应、批准工具以及恢复会话 |\n| [CCHub](https:\u002F\u002Fgithub.com\u002FMoresl\u002Fcchub) | - | 基于 Tauri 2 的桌面应用，用于管理 Claude Code 生态系统：MCP 市场、配置文件、技能、工作流、钩子、安全审计和自动驾驶功能。跨平台，约 20MB |\n| [Ruflo](https:\u002F\u002Fgithub.com\u002Fruvnet\u002Fruflo) | 21,200+ | 代理编排平台——部署多代理集群，协调自主工作流，实现分布式群体智能，并集成 RAG 功能 |\n| [Vibe Kanban](https:\u002F\u002Fgithub.com\u002FBloopAI\u002Fvibe-kanban) | 23,200+ | 基于看板的编排工具，适用于 10 多个编码代理，每个代理拥有独立的 Git 工作树 |\n| [Claude-Code-Workflow](https:\u002F\u002Fgithub.com\u002Fcatlog22\u002FClaude-Code-Workflow) | 1,400+ | 基于 JSON 的多代理协作框架，配备智能 CLI 编排功能（支持 Gemini\u002FQwen\u002FCodex） |\n| [ccswarm](https:\u002F\u002Fgithub.com\u002Fnwiizo\u002Fccswarm) | 127 | 基于 Rust 的多代理编排工具，提供专用代理池、Git 工作树隔离功能，并可减少 93% 的 Token 消耗 |\n| [parallel-code](https:\u002F\u002Fgithub.com\u002Fjohannesjo\u002Fparallel-code) | 370+ | 同时运行 Claude Code、Codex 和 Gemini，每个代理都在独立的 Git 工作树中执行 |\n| [Bernstein](https:\u002F\u002Fgithub.com\u002Fchernistry\u002Fbernstein) | 5+ | Python 多代理编排工具——在独立的 Git 工作树上并行启动 Claude Code、Codex CLI 和 Gemini CLI，通过测试验证后自动提交代码。协调过程中不消耗任何 LLM Token |\n| [Poirot](https:\u002F\u002Fgithub.com\u002FLeonardoCardoso\u002FPoirot) | 96 | macOS 应用程序，用于浏览 Claude Code 会话、查看差异并重新执行命令。读取本地日志文件，可在离线状态下运行 |\n| [TokenEater](https:\u002F\u002Fgithub.com\u002FAThevon\u002FTokenEater) | 179 | 原生 macOS 菜单栏应用程序，用于监控 Claude AI 使用限制并实时观看编码会话 |\n| [Claw](https:\u002F\u002Fgithub.com\u002Fjamesrochabrun\u002FClaw) | 86 | 用 Swift 封装 Claude Code SDK 的原生 macOS 应用程序。包含计划模式、MCP 集成和自定义系统提示功能 |\n| [The Claude Protocol](https:\u002F\u002Fgithub.com\u002FAvivK5498\u002FThe-Claude-Protocol) | 149 | 一套围绕 Claude Code 构建的强制执行层，配备 13 个钩子——可阻止不安全操作并强制实施工作树隔离 |\n| [Notch So Good](https:\u002F\u002Fgithub.com\u002Fdeepshal99\u002Fnotch-so-good) | 新 | 基于 macOS 通知中心的会话监控工具，配有像素风伴侣界面、13 种动画效果、智能通知以及多会话支持 |\n| [Anima](https:\u002F\u002Fgithub.com\u002Fbtangonan\u002Fanima) | 新 | Claude Code 的原生 macOS 伴侣应用。为每个项目生成 ASCII 形象伙伴，内置 NIM 代币经济系统，并配备跨会话监视守护进程。基于 Tauri v2 和 Rust 开发，二进制文件仅 4MB |\n| [crit](https:\u002F\u002Fgithub.com\u002Ftomasz-tomczyk\u002Fcrit) | 105 | 本地浏览器 UI，用于对任意文件或代理输出进行内联代码评审；可通过计划钩子与 Claude Code 集成，在执行前审查并批准计划，并输出结构化的 .crit.json 文件供代理使用。 |\n| [telegram-ai-bridge](https:\u002F\u002Fgithub.com\u002FAliceLJY\u002Ftelegram-ai-bridge) | 新 | 通过 Telegram 在手机上运行 N 个并行的 Claude Code 会话。具备作战室模式、针对不同机器人的个性化角色设定、A2A 多代理协作以及手机端工具审批功能。支持 Claude Code、Codex 和 Gemini 后端 |\n| [Untether](https:\u002F\u002Fgithub.com\u002Flittlebearapps\u002Funtether) | 12 | Claude Code 的 Telegram 桥接工具（还可连接其他 5 个 AI 代理）。可通过语音或文字发送任务，实时查看进度，并使用手机上的内联按钮批准更改 |\n| [OctoAlly](https:\u002F\u002Fgithub.com\u002Fai-genius-automations\u002Foctoally) | 18 | AI 编码会话编排仪表盘——通过 RuFlo 实现多代理协同思维，支持 Whisper 语音转录、tmux 会话持久化、内置 Git 版本控制，并采用 Electron 桌面应用 |\n| [tokburn](https:\u002F\u002Fgithub.com\u002Flsvishaal\u002Ftokburn) | 3 | Claude Code 的本地分析仪表盘——计算 API 等效成本，检测资源浪费模式（重复读取、徘徊不前、高成本异常），提供 Web UI 和 REST API。无需网络请求、无需账号、无需遥测数据。`uvx tokburn serve`。 |\n| [skill-builder](https:\u002F\u002Fgithub.com\u002FScottpedia0\u002Fskill-builder) | 新 | 分析用户活动（终端历史、Git、浏览器、Claude 对话记录），并通过 LLM 生成实用技能。内置 15 种技能，外加无限量的 AI 生成技能。配备 React UI 和 MCP 服务器构建工具 |\n| [claude-workspace-snapshot](https:\u002F\u002Fgithub.com\u002FREMvisual\u002Fclaude-workspace-snapshot) | 新 | 将正在运行的 Claude Code 会话快照并恢复为命名且带颜色标记的 Windows Terminal 标签页。通过进程检查和 .jsonl 文件活动检测正在运行的会话。仅限 Windows 平台 |\n| [KANBAII](https:\u002F\u002Fgithub.com\u002Fmartinmsaavedra\u002Fkanbaii) | 新 | 面向 Claude Code 的原生 AI 看板——可直观规划、跟踪进度并让 AI 自动执行。支持顺序引擎（Ralph）、并行多代理（Teams）、成本追踪以及实时仪表盘。`npx kanbaii start`。 |\n| [Claude Session Visualizer](https:\u002F\u002Fgithub.com\u002Fanaypaul\u002Fclaude-session-visualizer) | 新 | Claude Code 会话的实时可视化仪表盘，包含执行树、Token 成本跟踪、思考轨迹探索以及错误调试功能 |\n| [CCHub](https:\u002F\u002Fgithub.com\u002FMoresl\u002Fcchub) | 新 | 基于 Tauri 2.0 的桌面应用，用于管理完整的 Claude Code 生态系统——MCP 服务器、技能、钩子和配置文件。可自动扫描 Claude Code、Claude Desktop 和 Cursor 的配置文件 |\n| [AionUI](https:\u002F\u002Fgithub.com\u002FiOfficeAI\u002FAionUi) | 20,500+ | 免费的本地开源 24\u002F7 协作应用，适用于 Gemini CLI、Claude Code、Codex 和 OpenCode |\n| [chops](https:\u002F\u002Fgithub.com\u002FShpigford\u002Fchops) | 1,000+ | macOS 应用程序，用于浏览、编辑和管理 Claude Code、Cursor、Codex、Windsurf 和 Amp 中的技能 |\n| [ClaudeCode Launchpad CLI](https:\u002F\u002Fgithub.com\u002Fnoambrand\u002Fkivun-terminal) | 9 | Windows 和 macOS 上的两分钟安装程序，用于快速部署 Claude Code——自动安装 Node.js、Git 和 Claude Code；配备两行实时状态栏（模型、上下文占比、使用限制及重置计时器）；提供带有文件选择器的桌面快捷方式；可在任何 Windows 文件夹上右键选择“使用 ClaudeCode Launchpad CLI 打开”；可选浅蓝色主题；支持 Claude 标志设置 |\n| [clideck](https:\u002F\u002Fgithub.com\u002Frustykuntz\u002Fclideck) | 31 | 类似 WhatsApp 的仪表盘，可在同一个浏览器窗口中管理 AI 编码代理（Claude Code、Codex、Gemini CLI、OpenCode）。提供实时状态、会话恢复、代理间自动路由以及移动端远程访问功能 |\n| [WhereMyTokens](https:\u002F\u002Fgithub.com\u002Fjeongwookie\u002FWhereMyTokens) | 新 | Windows 系统托盘应用程序，用于实时监控 Claude Code 的 Token 使用情况。显示每会话的 Token 数量、成本、上下文窗口进度、速率限制条形图（5 小时\u002F1 周）、活动热力图以及模型细分。可注册为 Claude Code statusLine 插件，实现无轮询的实时数据同步。仅限 Windows 平台 |\n| [Onepilot](https:\u002F\u002Fonepilotapp.com) | - | iOS 应用程序，可通过 SSH 从手机远程连接到服务器，运行 Claude Code、Codex 等编码代理。内置 SwiftTerm 终端模拟器，支持代理会话管理和移动端访问开发机器 |\n| [docker-claude-code](https:\u002F\u002Fgithub.com\u002Fgw0\u002Fdocker-claude-code) | 新 | 在隔离的 Docker 容器中运行 Claude Code，支持多配置文件、安全加固、最佳实践默认设置，预装一系列插件\u002F技能包，并提供远程开发支持。可直接替换 `claude` 命令——只需一个简单的 shell 别名即可。 |\n| [amux](https:\u002F\u002Fgithub.com\u002Fmixpeek\u002Famux) | 新 | 开源代理多路复用器，可在 Web 仪表盘下同时运行数十个并行的 Claude Code 会话，具备自我修复的监控机制、看板、代理间 REST API 以及移动端 PWA。仅需一个 Python 文件 |\n| [cc-token-status](https:\u002F\u002Fgithub.com\u002Fechowonderfulworld\u002Fcc-token-status) | 新 | macOS 菜单栏中的 Claude Code 仪表盘。显示计划使用限制、成本、Token 数量、趋势、模型细分以及通过 iCloud 实现的多设备同步。仅需一个 Python 文件，可自动更新 |\n\n---\n\n\n\n## 生态系统\n\nClaude Code 生态系统中的知名项目、目录和资源。\n\n| 名称 | 星级 | 描述 |\n|------|-------|-------------|\n| [claude-mem](https:\u002F\u002Fgithub.com\u002Fthedotmack\u002Fclaude-mem) | 35,900+ | 自动捕获 Claude 的所有操作，通过 AI 压缩，并将上下文注入到未来的会话中。2026年2月 GitHub 趋势榜第一 |\n| [wshobson\u002Fagents](https:\u002F\u002Fgithub.com\u002Fwshobson\u002Fagents) | 31,300+ | 112 个专业代理、16 个编排器、146 种技能和 79 种工具，分布在 72 个专注插件中 |\n| [oh-my-claudecode](https:\u002F\u002Fgithub.com\u002FYeachan-Heo\u002Foh-my-claudecode) | 9,900+ | 面向团队的多代理编排系统，包含 19 个专业代理和 28 种技能 |\n| [ccusage](https:\u002F\u002Fgithub.com\u002Fryoppippi\u002Fccusage) | 11,500+ | 用于从本地 JSONL 文件分析 Claude Code 使用情况的命令行工具。支持离线模式，无需任何 API 调用 |\n| [cc-statistics](https:\u002F\u002Fgithub.com\u002FandroidZzT\u002Fcc-statistics) | 270+ | 三合一的 Claude Code 统计工具：命令行界面 + Web 界面 + macOS 原生 SwiftUI 面板。提供令牌成本、按语言划分的代码变更、效率评分及每周报告等功能。同时支持 Codex 和 Cursor |\n| [ccpm](https:\u002F\u002Fgithub.com\u002Fautomazeio\u002Fccpm) | 7,600+ | 结合 GitHub Issues 和 Git worktrees 的项目管理工具，支持并行代理执行 |\n| [claude-code-system-prompts](https:\u002F\u002Fgithub.com\u002FPiebald-AI\u002Fclaude-code-system-prompts) | 5,900+ | 提取自 Claude Code 的系统提示，并随每次发布更新 |\n| [claude-context](https:\u002F\u002Fgithub.com\u002Fzilliztech\u002Fclaude-context) | 5,600+ | Zilliz 推出的语义代码搜索 MCP——混合 BM25 和向量检索，可减少约 40% 的令牌用量 |\n| [claude-skills](https:\u002F\u002Fgithub.com\u002Falirezarezvani\u002Fclaude-skills) | 5,300+ | 涵盖 9 个领域的 192 种技能，配备 254 个 Python 自动化工具 |\n| [paco-framework](https:\u002F\u002Fgithub.com\u002FPenguinAlleyApps\u002Fpaco-framework) | 1+ | 以 Markdown 优先的多代理操作系统，专为 Claude Code 设计。协调 3 到 16 个专业 AI 代理（如工程、QA、增长、财务等），采用基于文件的调度机制、机构记忆、CEO 审批以及 24\u002F7 全天候排程功能。无需 Python。MIT 许可证 |\n| [peon-ping](https:\u002F\u002Fgithub.com\u002FPeonPing\u002Fpeon-ping) | 3,900+ | 为 Claude Code 及其他代理提供《魔兽争霸 III》农民语音通知 |\n| [n8n-skills](https:\u002F\u002Fgithub.com\u002Fczlonkowski\u002Fn8n-skills) | 3,400+ | 用于构建生产就绪 n8n 工作流的技能集 |\n| [claude-hooks-mastery](https:\u002F\u002Fgithub.com\u002Fdisler\u002Fclaude-code-hooks-mastery) | 3,300+ | 完整的 Claude Code 钩子精通指南，附带可直接投入生产的脚本 |\n| [claude-supermemory](https:\u002F\u002Fgithub.com\u002Fsupermemoryai\u002Fclaude-supermemory) | 2,300+ | 利用 Supermemory 实现跨会话的持久化内存 |\n| [myclaude](https:\u002F\u002Fgithub.com\u002Fstellarlinkco\u002Fmyclaude) | 2,400+ | 多代理编排路由至 Claude Code、Codex、Gemini 和 OpenCode |\n| [claude-code-mcp](https:\u002F\u002Fgithub.com\u002Fsteipete\u002Fclaude-code-mcp) | 1,100+ | 将 Claude Code 作为一次性 MCP 服务器运行——成为你代理中的一个代理 |\n| [ccmanager](https:\u002F\u002Fgithub.com\u002Fkbwo\u002Fccmanager) | 940+ | 支持 8 个编码代理的会话管理器，具备智能自动审批功能 |\n| [cog](https:\u002F\u002Fgithub.com\u002Fmarciopuga\u002Fcog) | 240+ | 适用于 Claude Code 的认知架构——通过纯文本约定实现持久化内存、自我反思和前瞻性思考。无依赖项，仅需 CLAUDE.md 和 Markdown 文件 |\n| [Cortex](https:\u002F\u002Fgithub.com\u002FSKULLFIRE07\u002Fcortex-memory) | —— | 用于编码助手的持久化 AI 内存。自动捕获决策、模式和上下文信息，贯穿各个会话。提供 VSCode 扩展、命令行工具和 MCP 服务器。免费 |\n| [openclaw-self-healing](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fopenclaw-self-healing) | 32+ | 针对 Claude Code 及任何服务的四层自主崩溃恢复系统——64% 的问题可自动解决，与 LLM 无关（Claude\u002FGPT-4\u002FGemini\u002FOllama），并提供 Prometheus 指标 |\n| [CCM](https:\u002F\u002Fgithub.com\u002Fdr5hn\u002Fccm) | 8+ | Claude Code 的高级用户工具包——支持多账户管理、项目绑定、会话清理、环境快照、权限审计和健康诊断。仅需一个 Bash 脚本，除 Bash 4.4+ 和 jq 外无其他运行时依赖 |\n| [openclaw-memorybox](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fopenclaw-memorybox) | 8+ | Claude Code 的内存卫生命令行工具——防止上下文溢出导致的崩溃，可将 MEMORY.md 文件大小减少 83%，且无任何依赖 |\n| [openclaw-self-evolving](https:\u002F\u002Fgithub.com\u002FRamsbaby\u002Fopenclaw-self-evolving) | 2+ | 每周自我改进流水线——扫描 Claude Code 日志，提出 CLAUDE.md\u002FAGENTS.md 规则修改建议，全程无需 API 费用 |\n| [caliber](https:\u002F\u002Fgithub.com\u002Fcaliber-ai-org\u002Fai-setup) | —— | 一款能够指纹识别项目并生成 AI 代理配置（CLAUDE.md、技能、AGENTS.md）的命令行工具。可评估质量、自动刷新，并支持 Claude Code、Cursor 和 Codex |\n| [faf-cli](https:\u002F\u002Fgithub.com\u002FWolfe-Jam\u002Ffaf-cli) | —— | AI 上下文的 package.json。IANA 注册的 `.faf` 格式——初始化、评分、与 `CLAUDE.md` 双向同步、导出至 AGENTS.md、GEMINI.md 和 .cursorrules。基于 Bun 构建 |\n| [claude-starter-kit](https:\u002F\u002Fgithub.com\u002Fawrshift\u002Fclaude-starter-kit) | 新 | 即用型项目结构，配备持久化内存、会话连续性、钩子以及 3 种捆绑技能（Gemini、头脑风暴、设计） |\n| [claude-code-kickstart](https:\u002F\u002Fgithub.com\u002Fypollak2\u002Fclaude-code-kickstart) | 新 | 主观性强的入门套件——只需一条命令即可安装精选的 MCP 服务器、钩子、代理和配置文件。内置自动检测功能，包含 12 个代理、10 个配置文件和 20 多条 Shell 命令 |\n| [claude-code-power-stack](https:\u002F\u002Fgithub.com\u002Fbluzername\u002Fclaude-code-power-stack) | 新 | 在一次安装中集成幽灵内存、对话搜索、会话命名以及 Manus 风格的规划功能，并附赠作弊表 PDF |\n| [clooks](https:\u002F\u002Fgithub.com\u002Fmauribadnights\u002Fclooks) | 新 | 持久化的钩子守护进程，取代每次调用时的启动方式——通过批处理、依赖解析和指标监控，使钩子速度提升 112 倍 |\n| [AIRIS MCP Gateway](https:\u002F\u002Fgithub.com\u002Fagiletec-inc\u002Fairis-mcp-gateway) | 新 | 基于 Docker 的 MCP 多路复用器，可将 60 多种工具整合到 7 个元工具背后，从而减少 97% 的上下文令牌使用量。只需一条命令即可启动，服务器会按需自动启用 |\n| [Claude Code AWS Gateway](https:\u002F\u002Fgithub.com\u002Fantkawam\u002Fclaude-code-aws-gateway) | 新 | 自托管的 API 网关，可通过 Amazon Bedrock 将 Claude Code 路由出去，支持团队 API 密钥、预算、速率限制、OIDC SSO、SCIM 配置以及用于分析的管理员门户 |\n| [gemini-claude-bridge](https:\u002F\u002Fgithub.com\u002Fweijiafu14\u002Fgemini-claude-bridge) | 新 | Gemini 至 Claude 的协议转换器，允许将 Gemini 模型用作 Claude Code 的后端。修复了 3 个 LiteLLM 错误 |\n| [spartan-ai-toolkit](https:\u002F\u002Fgithub.com\u002Fspartan-stratos\u002Fspartan-ai-toolkit) | 新 | 为 Claude Code 提供工程纪律层——包含 67 条斜杠命令、20 条编码规则、27 种技能、9 个代理以及每一步之间的质量门控。提供 8 种堆栈配置（Go、Python、Java、Kotlin、React 等），并支持跨会话的代理内存。安装命令：`npx @c0x12c\u002Fspartan-ai-toolkit@latest --local` |\n| [AgenTopology](https:\u002F\u002Fgithub.com\u002Fagentopology\u002Fagentopology) | 新 | 用于多代理编排的声明式语言和命令行工具——可在单个 `.at` 文件中定义代理、流程、门控、钩子和群聊，支持快速搭建至 7 个平台，并提供交互式可视化工具。Apache 2.0 许可证 |\n| [cognitropy-lab](https:\u002F\u002Fgithub.com\u002FDaxxSec\u002Fcognitropy-lab) | —— | 由日期种子熵引擎驱动的自主工作空间工厂（363 个领域、22 个类别、1880 万种独特结果）。每日构建一个符合 Agent Workspace Model 标准的新工作空间——从固件逆向工程到水耕栽培再到卫星通信——全程配备完整的斜杠命令、工作流和秘密扫描功能。完全无需人工干预。\n| [brood-box](https:\u002F\u002Fgithub.com\u002Fstacklok\u002Fbrood-box) | 10+ | 在硬件隔离的微型虚拟机中运行 AI 编码代理（Claude Code、Codex、OpenCode），实现快照隔离、出口控制和 MCP 授权 |\n| [claude-kit](https:\u002F\u002Fgithub.com\u002Fluiseiman\u002Fclaude-kit) | 1 | Claude Code 的配置工厂——包含 13 个可组合堆栈、6 个代理、审计评分和实践流水线。可在不同项目间引导并维护 .claude\u002F 目录 |\n| [Global Chat](https:\u002F\u002Fgithub.com\u002Fpumanitro\u002Fglobal-chat) | —— | 跨协议 MCP 服务器发现——可在 6 个以上注册表中搜索超过 18,000 个服务器。同时也支持 A2A 和 agents.txt。免费的 agents.txt 验证工具。NPM 包：`@global-chat\u002Fmcp-server` |\n| [GAAI Framework](https:\u002F\u002Fgithub.com\u002FFr-e-d\u002FGAAI-framework) | 95+ | 即插即用的治理层（.gaai\u002F 文件夹）——包括待办事项授权、跨会话记忆、决策追踪、质量门控以及自主交付守护进程。适用于 Claude Code、Cursor、Codex CLI 和 Gemini CLI |\n| [Wiggum CLI](https:\u002F\u002Fgithub.com\u002Ffedericoneri\u002Fwiggum-cli) | —— | 开源 AI 代理，可扫描代码库（80 多种技术），通过 AI 面谈生成规格，并利用 Claude Code 或 Codex 运行自主 Ralph 循环。代理模式可端到端地在 GitHub 上提交问题，具备优先级感知调度和自动合并功能 |\n| [expert-dispatch](https:\u002F\u002Fgithub.com\u002Fsimonsysun\u002Fexpert-dispatch) | 新 | 专家派遣模式——让低成本 AI 助手将复杂任务委托给 Claude Code CLI。提供 Bash 适配器，支持运行\u002F恢复\u002F审查的工作流、会话连续性、项目专用目录以及结构化的 JSON 输出。专为其他 AI 系统调用而设计 |\n| [Claude Command Center](https:\u002F\u002Fgithub.com\u002Ftuning-labs-oss\u002Fclaude-command-center) | —— | 经过生产验证的工作空间模板，配备会话钩子、结构化内存（4 种类型）、多领域路由、每日任务跟踪以及交互式 \u002Fsetup 入门流程。包含 6 种领域模板、预提交安全措施和示例技能 |\n| [Code Insights](https:\u002F\u002Fgithub.com\u002Fmelagiri\u002Fcode-insights) | 8 | 本地优先的命令行工具和仪表盘，用于分析来自 Claude Code、Cursor、Codex CLI、Copilot CLI 和 VS Code Copilot 的 AI 编码会话。基于 SQLite，提供终端分析、浏览器仪表盘以及 LLM 驱动的洞察 |\n| [claude-code-cheat-sheet](https:\u002F\u002Fcc.storyfox.cz\u002F) | 新 | 完整的一页可打印参考手册——涵盖所有快捷键、命令、CLI 标志、MCP 配置、内存、技能和代理。每日根据官方文档自动更新。[英文](https:\u002F\u002Fcc.storyfox.cz\u002F) · [中文](https:\u002F\u002Fcc.storyfox.cz\u002Fzh\u002F) · [日文](https:\u002F\u002Fcc.storyfox.cz\u002Fjp\u002F) · [韩文](https:\u002F\u002Fcc.storyfox.cz\u002Fkr\u002F) |\n| [agent-dotfiles](https:\u002F\u002Fgithub.com\u002Fsaqibameen\u002Fagent-dotfiles) | 新 | 编写一次 AI 编码规则，即可同步到所有代理。支持 Command Code、Claude Code、Cursor、Copilot、Codex 和 OpenCode |\n| [zclean](https:\u002F\u002Fgithub.com\u002Fwhynowlab\u002Fzclean) | 19+ | 杀死 Claude Code 和 Codex 留下的孤儿进程。阻止僵尸节点\u002FPython\u002Fesbuild 消耗你的 RAM |\n| [Hindsight](https:\u002F\u002Fgithub.com\u002Fvectorize-io\u002Fhindsight) | —— | Vectorize 推出的先进长期记忆系统，适用于 AI 代理。仿生式的保留、召回和反思功能，配备 4 种并行检索策略。可自托管或部署在云端，MIT 许可证。[Claude Code 集成](https:\u002F\u002Fhindsight.vectorize.io\u002Fintegrations) |\n| [healthcare-agents](https:\u002F\u002Fgithub.com\u002Fajhcs\u002Fhealthcare-agents) | 新 | 51 个专注于医疗保健管理的专业代理，具备 MHA 级别的专业知识，覆盖收入周期、合规、质量、临床运营、支付方关系、健康 IT 等 10 个部门 |\n| [cc-switch](https:\u002F\u002Fgithub.com\u002Ffarion1231\u002Fcc-switch) | 35,500+ | 跨平台桌面一体化助手，适用于 Claude Code、Codex、OpenCode 和 Gemini CLI |\n| [claude-code-router](https:\u002F\u002Fgithub.com\u002Fmusistudio\u002Fclaude-code-router) | 30,700+ | 将 Claude Code 用作编码基础设施基础，支持自定义模型路由和交互 |\n| [SuperClaude](https:\u002F\u002Fgithub.com\u002FSuperClaude-Org\u002FSuperClaude_Framework) | 22,000+ | 配置框架，包含专业命令、认知角色和开发方法论 |\n| [claude-code-templates](https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates) | 23,800+ | 用于配置和监控 Claude Code 项目及工作流的命令行工具 |\n| [planning-with-files](https:\u002F\u002Fgithub.com\u002FOthmanAdi\u002Fplanning-with-files) | 17,600+ | 类似 Manus 的持久化 Markdown 规划技能，用于结构化项目管理 |\n| [claude-hud](https:\u002F\u002Fgithub.com\u002Fjarrodwatts\u002Fclaude-hud) | 15,200+ | 插件形式显示上下文使用情况、当前工具、正在运行的代理以及待办事项进度的 HUD 叠加层 |\n| [codachi](https:\u002F\u002Fgithub.com\u002Fvincent-k2026\u002Fcodachi) | 新 | 像电子宠物一样的状态栏小生物，会随着上下文使用量的增长而成长，同时显示缓存命中率和资源消耗速度 |\n| [awesome-claude-code-subagents](https:\u002F\u002Fgithub.com\u002FVoltAgent\u002Fawesome-claude-code-subagents) | 15,600+ | 按领域分类的 100 多个专业 Claude Code 子代理 |\n| [serena](https:\u002F\u002Fgithub.com\u002Foraios\u002Fserena) | 22,200+ | 用于编码代理的语义检索和编辑 MCP 服务器——具备代码感知的搜索和导航功能 |\n| [Understand-Anything](https:\u002F\u002Fgithub.com\u002FLum1104\u002FUnderstand-Anything) | 7,000+ | 将任何代码库转化为可交互的知识图谱，便于探索 |\n| [n8n-mcp](https:\u002F\u002Fgithub.com\u002Fczlonkowski\u002Fn8n-mcp) | 17,000+ | 专为 Claude Code 设计的 MCP，用于构建和管理 n8n 自动化工作流 |\n| [ToolRouter](https:\u002F\u002Ftoolrouter.com) | —— | 为 Claude Code 赋予超能力——只需一个账号即可按需使用 150 多种工具。包括竞争对手研究、视频制作、网页搜索、图像生成、安全扫描等。`claude mcp add toolrouter -- npx -y toolrouter-mcp` |\n| [Claudoscope](https:\u002F\u002Fgithub.com\u002Fcordwainersmith\u002Fclaudoscope) | 29 | macOS 原生菜单栏应用，用于 Claude Code 会话分析。可跟踪令牌和成本、进行全文搜索、实时检测秘密、检查配置健康状况（44 条规则）。读取本地 JSONL 文件，完全离线，零遥测数据。采用 Swift\u002FSwiftUI 开发，可通过 Homebrew 安装 |\n| [LynxPrompt](https:\u002F\u002Fgithub.com\u002FGeiserX\u002FLynxPrompt) | 25+ | 开源平台，用于管理 AI 编码代理的配置文件（CLAUDE.md、AGENTS.md、.cursorrules、copilot-instructions.md）。提供 Web 市场、命令行工具、VS Code 扩展，并支持通过 Docker\u002FHelm 自主托管及联邦协作 |\n\n---\n\n\n\n## XVARY 股票研究\n\n- [XVARY 股票研究](https:\u002F\u002Fgithub.com\u002Fxvary-research\u002Fclaude-code-stock-analysis-skill) — 用于公开 SEC EDGAR 数据及市场数据的 Claude Code 技能：`\u002Fanalyze`、`\u002Fscore`、`\u002Fcompare`。MIT 许可。\n\n---\n\n## 资源\n\n关于 Claude Code 的教程、指南和深度解析。\n\n| 资源 | 类型 | 描述 |\n|----------|------|-------------|\n| [Claude Code 官方文档](https:\u002F\u002Fdocs.anthropic.com\u002Fen\u002Fdocs\u002Fclaude-code) | 文档 | Anthropic 官方发布的 Claude Code 文档 |\n| [上下文工程指南](https:\u002F\u002Fwww.anthropic.com\u002Fresearch\u002Flong-running-Claude) | 指南 | Anthropic 提供的关于长时间运行 Claude Code 会话及科学计算的指南 |\n| [自动模式工程](https:\u002F\u002Fwww.anthropic.com\u002Fengineering\u002Fclaude-code-auto-mode) | 博客 | 自动模式的双层安全系统如何在底层运作 |\n| [Claude Code 指南](https:\u002F\u002Fclaudecodeguide.dev) | 指南 | 零术语、适合初学者的指南，从首次安装到日常使用系统（[GitHub](https:\u002F\u002Fgithub.com\u002Fmshadmanrahman\u002Fclaudecode-guide)） |\n| [Claude Code 钩子精通](https:\u002F\u002Fgithub.com\u002Fdisler\u002Fclaude-code-hooks-mastery) | 教程 | 包含生产级钩子脚本的完整精通指南 |\n| [技能、MCP 与插件的区别](https:\u002F\u002Fwww.morphllm.com\u002Fcontext-engineering) | 指南 | 在 Claude Code 中何时使用哪种扩展机制 |\n| [上下文工程入门](https:\u002F\u002Fnewsletter.victordibia.com\u002Fp\u002Fcontext-engineering-101-how-agents) | 文章 | 三种核心策略：压缩、隔离、智能体记忆 |\n| [智能体团队详解](https:\u002F\u002Fwww.turingcollege.com\u002Fblog\u002Fclaude-agent-teams-explained) | 指南 | 多个 Claude Code 实例如何协作组成 AI 工程团队 |\n| [84 条最佳实践](https:\u002F\u002Fdiscuss.huggingface.co\u002Ft\u002F10-essential-claude-code-best-practices-you-need-to-know\u002F174731) | 收集 | Hugging Face 社区整理的 Claude Code 最佳实践 |\n| [通过命令行控制电脑](https:\u002F\u002Fcode.claude.com\u002Fdocs\u002Fen\u002Fcomputer-use) | 文档 | 让 Claude 控制你的桌面——打开应用、点击、输入、截图。支持 macOS、Pro\u002FMax，v2.1.85+。需逐个应用批准、分层访问，并可通过 Esc 键中断 |\n| [调度与远程控制](https:\u002F\u002Fclaude.com\u002Fblog\u002Fdispatch-and-computer-use) | 博客 | 可通过手机或网页运行 Claude Code，安排定期任务，远程控制会话 |\n| [Claude Code 备忘单](https:\u002F\u002Fcc.storyfox.cz\u002F) | 参考 | 一页可打印的每日自动更新参考，支持四种语言 |\n| [多智能体编排](https:\u002F\u002Faddyosmani.com\u002Fblog\u002Fcode-agent-orchestra\u002F) | 博客 | Addy Osmani 对多智能体编码模式的综述 |\n\n## 相关优秀列表\n\n| 列表 | 星数 | 重点 |\n|------|-------|-------|\n| [awesome-mcp-servers](https:\u002F\u002Fgithub.com\u002Fappcypher\u002Fawesome-mcp-servers) | 5,300+ | 精选的 MCP 服务器，涵盖 30 个类别 |\n| [awesome-ai-agents](https:\u002F\u002Fgithub.com\u002Fe2b-dev\u002Fawesome-ai-agents) | 27,000+ | 开源与闭源的 AI 智能体 |\n| [awesome-llm-apps](https:\u002F\u002Fgithub.com\u002FShubhamsaboo\u002Fawesome-llm-apps) | 104,000+ | 基于 LLM 的应用及多智能体系统 |\n| [awesome-claude-code (hesreallyhim)](https:\u002F\u002Fgithub.com\u002Fhesreallyhim\u002Fawesome-claude-code) | 34,000+ | 技能、钩子、斜杠命令以及编排工具 |\n| [antigravity-awesome-skills](https:\u002F\u002Fgithub.com\u002Fsickn33\u002Fantigravity-awesome-skills) | 28,700+ | 1,326+ 可通过 CLI 安装的智能体技能 |\n| [awesome-claude-code-subagents](https:\u002F\u002Fgithub.com\u002FVoltAgent\u002Fawesome-claude-code-subagents) | 15,600+ | 100+ 专门化的 Claude Code 子智能体 |\n| [claude-code-best-practice](https:\u002F\u002Fgithub.com\u002Fshanraisshan\u002Fclaude-code-best-practice) | 25,000+ | 全面的 Claude Code 配置参考实现 |\n| [awesome-claude-skills (Composio)](https:\u002F\u002Fgithub.com\u002FComposioHQ\u002Fawesome-claude-skills) | 49,300+ | 30 个精选技能 + 通过 Composio 提供的 832 个 SaaS 自动化模板。尤其擅长文档处理、创意及商业相关技能 |\n\n## 贡献\n\n欢迎贡献。请参阅 [CONTRIBUTING.md](CONTRIBUTING.md) 获取指南。\n\n## 许可证\n\n[Apache-2.0](LICENSE)\n| [AutoSearch](https:\u002F\u002Fgithub.com\u002F0xmariowu\u002FAutosearch) | 新 | 适用于 Claude Code 的自我进化型深度研究插件——32 个专用搜索频道（arXiv、GitHub、Reddit、HN、知乎、哔哩哔哩、CSDN 等 25 个），LLM 评分评估，以 Markdown\u002FHTML\u002F幻灯片形式呈现引用报告，跨会话学习。无需任何 API 密钥 |","# awesome-claude-code-toolkit 快速上手指南\n\n`awesome-claude-code-toolkit` 是一个专为 **Claude Code** 打造的综合工具集，汇集了 176+ 插件、135 个智能体（Agents）、42 条命令及大量实用技能与规则，旨在极大扩展 Claude Code 在开发工作流中的能力。\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\n*   **操作系统**：macOS, Linux 或 Windows (WSL2 推荐)。\n*   **核心依赖**：\n    *   已安装并配置好 **Claude Code** CLI 工具。\n    *   **Git**：用于克隆仓库或管理版本。\n    *   **Node.js & npm** (可选但推荐)：部分插件和脚本依赖 Node 环境。\n    *   **Bash\u002FZsh**：大多数安装脚本和 Hooks 基于 Shell 运行。\n*   **网络环境**：由于资源托管于 GitHub，国内用户建议配置代理或使用加速镜像以确保下载速度。\n\n## 安装步骤\n\n你可以通过以下三种方式之一进行安装，推荐使用 **一键安装脚本** 或 **插件市场** 方式。\n\n### 方式一：一键安装脚本（推荐）\n\n这是最快捷的方式，会自动处理目录结构和基础配置。\n\n```bash\ncurl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fmain\u002Fsetup\u002Finstall.sh | bash\n```\n\n> **提示**：若下载速度慢，可尝试先将 `raw.githubusercontent.com` 替换为国内加速地址（如 `ghproxy.net` 等），或在终端开启代理后执行。\n\n### 方式二：通过插件市场安装\n\n如果你已在 Claude Code 环境中，可直接使用内置命令添加：\n\n```bash\n\u002Fplugin marketplace add rohitg00\u002Fawesome-claude-code-toolkit\n```\n\n### 方式三：手动克隆\n\n适合需要自定义路径或离线部署的高级用户：\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit.git ~\u002F.claude\u002Fplugins\u002Fclaude-code-toolkit\n```\n\n## 基本使用\n\n安装完成后，工具集会自动加载相关的插件、技能和规则。以下是几种最简单的使用场景：\n\n### 1. 调用特定领域插件\n\n直接在 Claude Code 对话中请求使用已安装的插件功能。例如，使用 `aws-cost-saver` 优化成本：\n\n```text\n请运行 aws-cost-saver 插件扫描当前项目的 AWS 资源配置，并提供优化建议。\n```\n\n或者使用 `a11y-audit` 进行无障碍检查：\n\n```text\n对项目前端代码执行 a11y-audit 插件，检查 WCAG 合规性。\n```\n\n### 2. 使用预设技能 (Skills)\n\n工具集包含了大量预定义的技能。你可以直接指示 Claude 激活特定技能来处理任务。例如，使用成本优化技能：\n\n```bash\nnpx skills add Sagargupta16\u002Fclaude-cost-optimizer\n```\n然后在对话中：\n```text\n启用 cost-optimizer 技能，分析当前的 Token 使用情况并给出节省方案。\n```\n\n### 3. 利用自动化 Hooks\n\n许多插件（如 `pro-workflow` 或 `cc-safe-setup`）安装了自动运行的 Hooks。当你执行特定操作（如提交代码、创建新会话）时，它们会自动触发。\n\n例如，安装安全防护钩子：\n```bash\nnpx cc-safe-setup\n```\n此后，当你尝试执行危险命令时，系统会自动拦截并提示确认，无需额外指令。\n\n### 4. 查看可用命令列表\n\n在 Claude Code 会话中，你可以询问当前加载的工具概览：\n\n```text\n列出当前可用的所有插件和命令。\n```\n\n通过以上步骤，你即可快速利用 `awesome-claude-code-toolkit` 增强你的 AI 编程工作流。更多高级用法（如多智能体编排、MCP 配置等）可查阅项目仓库中的详细文档。","某初创公司后端团队需要在三天内重构遗留的 AWS 云架构，同时严格控制成本并保证零停机部署。\n\n### 没有 awesome-claude-code-toolkit 时\n- **成本失控风险**：工程师需手动编写脚本检查资源闲置情况，容易遗漏隐藏的高费实例，导致预算超支。\n- **协作流程混乱**：缺乏标准化的 Agent 分工，代码审查、测试和发布由同一提示词触发，常出现上下文丢失或任务冲突。\n- **环境配置繁琐**：每次新项目都要重复配置 MCP 连接和本地钩子（Hooks），耗时且容易因版本差异出错。\n- **技能管理无序**：团队积累的自定义技能（Skills）散落在不同文件中，无法统一审计或去重，复用率极低。\n\n### 使用 awesome-claude-code-toolkit 后\n- **智能成本优化**：直接调用 `aws-cost-saver` 插件，通过 173 项自动化检查瞬间识别冗余资源，实现 ML 驱动的规格调整，预计降低 60% 云成本。\n- **角色化代理协作**：引入 `gstack` 插件，让 Claude 分别扮演 CEO、工程经理和 QA 专家，自动并行处理需求分析、代码生成与质量验收，流程井然有序。\n- **一键标准化落地**：利用内置的 14 种 MCP 配置模板和 20 个预置 Hooks，分钟级完成开发环境搭建，确保全员工具链一致。\n- **技能资产治理**：使用 `skills-janitor` 自动审计并清理重复技能，将团队分散的经验沉淀为可追踪、可复用的标准资产库。\n\nawesome-claude-code-toolkit 将原本碎片化的 AI 辅助开发转变为具备企业级治理能力的自动化流水线，让团队从“配置工具”转向“专注交付”。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Frohitg00_awesome-claude-code-toolkit_ea3c2c74.png","rohitg00","Rohit Ghumare","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Frohitg00_0db5b2cc.png","Principal Product Evangelist | DevRel | GDE @Google Cloud | CNCF Ambassador | Docker Captain | AWS CommunityBuilder | Ex - Solo.io, Cerbos, Oracle, Reliance Jio","iii.dev","London, UK",null,"ghumare64","https:\u002F\u002Fwww.devrelasservice.com\u002F","https:\u002F\u002Fgithub.com\u002Frohitg00",[86,90,94],{"name":87,"color":88,"percentage":89},"JavaScript","#f1e05a",50.7,{"name":91,"color":92,"percentage":93},"Python","#3572A5",39.7,{"name":95,"color":96,"percentage":97},"Shell","#89e051",9.6,1285,383,"2026-04-16T06:35:08","Apache-2.0","Linux, macOS, Windows","未说明",{"notes":105,"python":103,"dependencies":106},"该工具是 Claude Code 的扩展工具集（包含插件、代理、技能等），而非独立的 AI 模型。运行前提是已安装并配置好 Claude Code 环境。安装方式支持通过插件市场、Git 克隆或一键脚本。部分特定插件可能需要 Node.js、Go 或其他语言运行时，具体取决于所选用的插件功能。",[107,108,109,110,111],"git","curl","bash","nodejs (部分插件需要)","npx (部分插件需要)",[14],[114,115,116,117,118,119],"claude","claude-code","claudecode","claudecode-hooks","plugins","skills","2026-03-27T02:49:30.150509","2026-04-17T09:53:25.526020",[123,128,133,138,143,148],{"id":124,"question_zh":125,"answer_zh":126,"source_url":127},36395,"如何向该列表提交新的技能或工具？","维护者建议直接提交 Pull Request (PR)。请将新条目添加到相应的表格中（例如 Community Skills 表），格式应包含：工具名称链接、安装命令（如 `npx skillkit install ...`）以及简短的功能描述。确保项目是真实的开源项目且具有实质性内容。","https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fissues\u002F7",{"id":129,"question_zh":130,"answer_zh":131,"source_url":132},36396,"为什么我的商业产品或付费工具提交被拒绝了？","该列表专注于开源和免费工具。如果提交的资源链接指向付费产品（如 Gumroad 商品）或包含模糊营销内容的商业网站，通常会被拒绝。请确保提交的项目是免费的且代码开源。","https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fissues\u002F180",{"id":134,"question_zh":135,"answer_zh":136,"source_url":137},36397,"什么样的项目描述会导致提交被关闭？","如果项目描述包含虚假信息（例如引用不存在的“泄露源码”）、过于夸张的营销用语而缺乏实质技术内容，或者仅仅是推广性文章而非标准的工具条目介绍，提交将会被关闭。请提供事实性的技术描述。","https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fissues\u002F179",{"id":139,"question_zh":140,"answer_zh":141,"source_url":142},36398,"如果我提交了重复的工具或已有 PR 在处理，会发生什么？","如果某个工具的提交与现有的 Issue 或已合并的 Pull Request 重复，新的 Issue 将会作为“重复项”被关闭。在提交前，请先搜索列表中是否已存在该工具或是否有正在进行的讨论。","https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fissues\u002F237",{"id":144,"question_zh":145,"answer_zh":146,"source_url":147},36399,"如何正确更新现有工具的条目信息？","如果需要更新现有条目（如版本号、下载量统计等），不应发布长篇的宣传性更新日志。维护者建议针对单个工具提交简洁的单行条目更新 PR，遵循列表中现有的格式，避免将其变成促销内容。","https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fissues\u002F212",{"id":149,"question_zh":150,"answer_zh":151,"source_url":152},36400,"提交的项目需要满足哪些基本质量标准？","项目应当是真实的代码仓库（如 GitHub  repo），具有一定的社区关注度（如 Star 数），并且有明确的技术用途。对于纯概念验证、缺乏实质代码或仅包含模糊营销文案的项目，通常不予收录。","https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fissues\u002F2",[154,159,164,169],{"id":155,"version":156,"summary_zh":157,"released_at":158},289221,"v0.4.0","## v0.4.0 新增内容\n\n### 合并了46个社区PR\n合并了所有待处理的社区贡献——每位贡献者都在 Git 历史中获得了应有的署名。针对11个无法自动合并的 PR，在本地解决了合并冲突。\n\n### 新插件（16个）\n- **ORCH** — 用于 AI 代理团队的 CLI 编排工具，配备状态机和 TUI 仪表盘\n- **AgentLint** — 33项检查的 AI 代理兼容性 Linter 工具\n- **AutoSearch** — 具有32个搜索渠道的自进化深度研究工具\n- **brooks-lint** — 基于经典软件工程书籍的 AI 代码评审工具（涵盖6本经典SE书籍）\n- **Claudebase** — 将 Claude Code 配置备份\u002F恢复\u002F同步到私有 GitHub 仓库\n- **CLIRank** — 对210+个 API 进行 CLI 友好度评分的 API 发现工具\n- **codebase-graph + agntk** — 结合 FalkorDB 知识图谱与 CLI AI 代理的 AST 解析工具\n- **Harness Evolver** — 原生集成 LangSmith 的自主代理优化工具\n- **\u002Fcht** — 提供13条斜杠命令和防崩溃自动保存功能的聊天持久化工具\n- **sitemd** — 使用22种 MCP 工具从 Markdown 构建网站\n- **knowledge-graph** — 适用于 Claude Code 的零依赖、原生 Git 内存层\n- **Leapfrog** — 多会话浏览器 MCP，支持15个并行 Chromium 会话\n- **2389 Research** — 包含4个插件的研究工具集（多智能体系统、评审小组、规范驱动开发、测试架构师）\n- **nv:context** — 跨6个杠杆维度的上下文工程技能\n\n### 新配套应用（9个）\n- **Claudoscope** — macOS 原生菜单栏会话分析工具（Swift\u002FSwiftUI）\n- **clideck** — 类似 WhatsApp 的仪表盘，用于管理 AI 编码代理\n- **WhereMyTokens** — Windows 系统托盘中的 Token 使用监控器\n- **ClaudeCode Launchpad CLI** — 2分钟完成的 Windows\u002FmacOS 安装程序\n- **Onepilot** — 通过 SSH 使用 Claude Code 的 iOS 应用\n- **Anima** — 基于 ASCII 的 macOS 伙伴应用，内置 NIM 代币经济体系（Tauri v2）\n- **telegram-ai-bridge** — 通过 Telegram 运行N个并行的 Claude Code 会话\n- **cc-token-status** — macOS 菜单栏使用情况仪表盘\n- **docker-claude-code** — 在隔离的 Docker 容器中运行 Claude Code\n\n### 新生态系统工具（8个）\n- **PA·co Framework** — 协调3至16个 AI 代理的多智能体操作系统\n- **Claude Code AWS Gateway** — 通过 Amazon Bedrock 自托管的 API 网关\n- **ToolRouter** — 使用一个账号按需访问150+种工具\n- **cognitropy-lab** — 自主生成每日工作空间的工具\n- **expert-dispatch** — Claude Code CLI 的专家派遣模式\n- **LynxPrompt** — AI 编码代理配置管理平台\n- **codachi** — 打造一只养成类的状态栏宠物\n- **faf-cli** — 经 IANA 注册的 AI 上下文格式 CLI 工具\n\n### 新社区技能（7个）\n- **AuraKit** — 涵盖37种模式的全栈开发技能\n- **cc-inspect** — 浏览器仪表盘，用于查看已安装的技能\u002F插件\u002F钩子\n- **floom** — 将 Python 脚本部署为云端自动化任务\n- **Obsidian 主题设计师** — 可视化的 Obsidian 主题设计工具\n- **SciAgent-Skills** — 197项生命科学相关技能（基因组学、蛋白质组学、药物研发等）\n- **manage-skills** — 从终端跨11种工具管理技能\n- **TemplateClaw** — 涵盖6大类的32个生产就绪模板\n\n### 新 MCP 配置\n- **Finance** — 图表库模式","2026-04-09T21:05:05",{"id":160,"version":161,"summary_zh":162,"released_at":163},289222,"v0.3.0","## v0.3.0 新增内容\n\n### 合并了45+ 个社区 PR\n逐一手动合并了所有待处理的社区贡献（每位贡献者都会在 Git 历史中获得署名）。对于 GitHub 上无法自动合并的 PR，我们在本地解决了合并冲突。\n\n### 新增12款高星工具\n- **cc-switch** (3.5万) — 跨平台的一体化 Claude Code\u002FCodex\u002FOpenCode\u002FGemini 工具\n- **claude-code-router** (3万) — 针对 Claude Code 的自定义模型路由工具\n- **SuperClaude** (2.2万) — 带认知角色的配置框架\n- **claude-code-templates** (2.3万) — 用于配置 Claude Code 项目的命令行工具\n- **serena** (2.2万) — 面向编码代理的语义检索 MCP 服务器\n- **AionUI** (2万) — 免费的 24\u002F7 协作应用，支持4个编码代理\n- **planning-with-files** (1.7万) — 类似 Manus 的持久化规划工具\n- **n8n-mcp** (1.7万) — 用于构建 n8n 工作流的 MCP\n- **claude-hud** (1.5万) — 上下文\u002F工具\u002F代理 HUD 叠加层\n- **awesome-claude-code-subagents** (1.5万) — 100 多种专业子代理\n- **Understand-Anything** (7千) — 代码库到知识图谱的转换工具\n- **chops** (1千) — 横跨5个代理的 macOS 技能管理器\n\n### 新增板块\n- **资源** — 10 篇教程、指南和深度解析（Anthropic 文档、上下文工程、代理团队、最佳实践）\n- **相关 Awesome 列表** — 7 个精选链接，指向 awesome-mcp-servers、awesome-ai-agents、awesome-llm-apps 等\n\n### 新增4个 MCP 配置\n- `security.json` — Ghidra 逆向工程、Snyk 漏洞扫描\n- `design.json` — Figma 设计上下文、Blender 3D 自动化\n- `workflow-automation.json` — n8n、Pipedream 集成\n- `mobile.json` — Android ADB、Xcode 构建工具\n\n### 星级历史图表\n在 README 页眉中嵌入了 star-history.com 图表，并支持深色\u002F浅色模式。\n\n### 更新统计信息\n- **176+ 插件** | **135 个代理** | **35 个精选 + 28 个社区技能**\n- **42 条命令** | **26 个配套应用** | **51 个生态条目**\n- **13 个 MCP 配置** | **15 条规则** | **7 个模板** | **5 个上下文**\n- **0 个未合并 PR，0 个未解决的问题**\n\n### 日常维护\n- 关闭了3个带有评论的重复 PR\n- 关闭了 issue #51（通过 PR #104 修复 marketplace.json 模式问题）\n- 关闭了 issue #159（营销请求）\n- 所有社区 PR 均已单独合并，以确保每位贡献者的署名权\n\n## 完整变更日志\nhttps:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit\u002Fcompare\u002Fv0.2.0...v0.3.0","2026-03-30T10:13:36",{"id":165,"version":166,"summary_zh":167,"released_at":168},289223,"v0.2.0","## 新增内容\n\n### 精选插件\n- **[pro-workflow](https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fpro-workflow)** — 经过实战检验的 Claude Code 工作流（1,400+ 颗星）\n- **[everything-claude-code](https:\u002F\u002Fgithub.com\u002Faffaan-m\u002Feverything-claude-code)** — 代理框架性能优化（78,600+ 颗星）\n- **[gstack](https:\u002F\u002Fgithub.com\u002Fgarrytan\u002Fgstack)** — Garry Tan 的 Claude Code 完整配置（15,000+ 颗星）\n\n### 新增 25+ 款生态工具\n高关注度工具现已列入工具箱：\n- **claude-mem**（35,900+ 颗星） — 自动捕获 + AI 压缩会话记忆\n- **wshobson\u002Fagents**（31,300+ 颗星） — 112 个专用代理，16 个编排器\n- **Vibe Kanban**（23,200+ 颗星） — 面向 10+ 编程代理的看板编排工具\n- **Opcode**（21,000+ 颗星） — 用于 Claude Code 的 Tauri 2 桌面 GUI\n- **Ruflo**（21,200+ 颗星） — 多代理集群编排平台\n- **claude-code-best-practice**（17,400+ 颗星） — 全面参考实现\n- **ccusage**（11,500+ 颗星） — 用于分析本地 JSONL 文件中 Claude Code 使用情况的命令行工具\n- **oh-my-claudecode**（9,900+ 颗星） — 以团队为中心的多代理编排工具\n- **CloudCLI**（8,400+ 颗星） — 免费的 Web\u002F移动端 Claude Code UI\n- **ccpm**（7,600+ 颗星） — 结合 GitHub Issues 和 Git 工作树的项目管理工具\n- 以及另外 15+ 款工具，每款拥有 100–6,000 颗星不等\n\n### 新增板块\n- **配套应用与 GUI** — 12 款适用于 Claude Code 的桌面\u002F网页\u002F移动端应用\n- **生态系统** — Claude Code 生态系统中的 15 个知名项目\n\n### 合并社区 PR 共 16 个\nonWatch、claude-recap、Poirot、avoid-ai-writing、Bouncer、dna-claude-analysis、claude-cost-optimizer、BGPT MCP、LightCMS、production-grade、Crypto\u002FDeFi MCP、MUSE、product-manager-skills、x-twitter-scraper、DiscoClaw、PUIUX Pilot\n\n### 关闭垃圾\u002F营销 PR 共 9 个\n清理了竞争对手链接替换、商业 SaaS 营销、失效链接、SEO 农场以及批量自我推广等内容。\n\n### 更新内容\n- SkillKit 市场：**400,000+ 技能**（从 15,000 提升）\n- 插件数量：**150+**（从 121 增加）\n- MCP 配置：**8**（从 7 增加）\n- 最后更新标识：**2026 年 3 月**\n\n## 数据统计\n- 150+ 款插件、135 个代理、35 个精选技能（通过 SkillKit 增加至 40 万）、42 条命令、19 个钩子、15 条规则、7 个模板、8 种 MCP 配置","2026-03-16T08:28:33",{"id":170,"version":171,"summary_zh":172,"released_at":173},289224,"v0.1.0","# 令人惊叹的 Claude 代码工具包 v0.1.0\n\n专为 Claude Code 打造的最全面工具包。\n\n## 包含内容\n\n| 类别       | 数量 |\n|------------|------|\n| 代理       | 135  |\n| 插件       | 120  |\n| 技能       | 35   (+通过 SkillKit 增加 15,000 个) |\n| 命令       | 42   |\n| 钩子       | 19 个脚本 |\n| 规则       | 15   |\n| 模板       | 7    |\n| MCP 配置   | 6    |\n| **文件总数** | **796** |\n\n## 安装\n\n```bash\n\u002Fplugin marketplace add rohitg00\u002Fawesome-claude-code-toolkit\n```\n\n或手动克隆：\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Frohitg00\u002Fawesome-claude-code-toolkit.git ~\u002F.claude\u002Fplugins\u002Fclaude-code-toolkit\n```\n\n## 亮点\n\n- **135 个代理**，涵盖 10 大类别（核心开发、语言专家、基础设施、QA、数据与 AI、开发者体验、专业领域、商业产品、编排、研究分析）\n- **120 个插件**，覆盖从智能提交到 Kubernetes 辅助工具等方方面面\n- 所有代理均使用 `model: opus`，以获得最佳性能\n- 集成 SkillKit，可通过 [agenstskills.com](https:\u002F\u002Fagenstskills.com) 获得超过 15,000 个附加技能\n\n## 许可证\n\nApache-2.0","2026-02-04T21:19:20"]