[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-IvanMurzak--Unity-MCP":3,"tool-IvanMurzak--Unity-MCP":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",140436,2,"2026-04-05T23:32:43",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":70,"readme_en":71,"readme_zh":72,"quickstart_zh":73,"use_case_zh":74,"hero_image_url":75,"owner_login":76,"owner_name":77,"owner_avatar_url":78,"owner_bio":79,"owner_company":80,"owner_location":81,"owner_email":82,"owner_twitter":76,"owner_website":83,"owner_url":84,"languages":85,"stars":116,"forks":117,"last_commit_at":118,"license":119,"difficulty_score":23,"env_os":120,"env_gpu":121,"env_ram":122,"env_deps":123,"category_tags":130,"github_topics":131,"view_count":23,"oss_zip_url":151,"oss_zip_packed_at":151,"status":16,"created_at":152,"updated_at":153,"faqs":154,"releases":185},2280,"IvanMurzak\u002FUnity-MCP","Unity-MCP","AI Skills, MCP Tools, and CLI for Unity Engine. Full AI develop and test loop. Use cli for quick setup. Efficient token usage, advanced tools. Any C# method may be turned into a tool by a single line. Works with Claude Code, Gemini, Copilot, Cursor and any other absolutely for free.","Unity-MCP 是一款专为 Unity 引擎打造的 AI 开发辅助工具，旨在打通从代码编写到测试验证的完整智能开发闭环。它基于模型上下文协议（MCP）构建，让开发者能够轻松地将任意 C# 方法转化为 AI 可调用的工具，只需一行代码即可实现。\n\n对于 Unity 开发者而言，手动编写重复性脚本或调试复杂逻辑往往耗时费力。Unity-MCP 通过提供高效的令牌（Token）使用策略和强大的命令行接口（CLI），显著降低了 AI 介入游戏开发的门槛。它不仅支持快速环境搭建，还能让 Claude Code、Gemini、GitHub Copilot、Cursor 等主流 AI 编程助手直接“理解”并操作 Unity 项目，从而自动化完成场景构建、逻辑实现及测试任务。\n\n这款工具特别适合 Unity 游戏开发者、技术美术以及希望提升编码效率的独立创作者。其核心亮点在于极高的兼容性与灵活性：既支持在 Unity 编辑器中运行，也适用于运行时环境；同时完全免费且开放，无缝对接各类大模型客户端。无论是资深程序员还是正在学习游戏开发的新手，都能借助 Unity-MCP 让 AI 成为得力的结对编程","Unity-MCP 是一款专为 Unity 引擎打造的 AI 开发辅助工具，旨在打通从代码编写到测试验证的完整智能开发闭环。它基于模型上下文协议（MCP）构建，让开发者能够轻松地将任意 C# 方法转化为 AI 可调用的工具，只需一行代码即可实现。\n\n对于 Unity 开发者而言，手动编写重复性脚本或调试复杂逻辑往往耗时费力。Unity-MCP 通过提供高效的令牌（Token）使用策略和强大的命令行接口（CLI），显著降低了 AI 介入游戏开发的门槛。它不仅支持快速环境搭建，还能让 Claude Code、Gemini、GitHub Copilot、Cursor 等主流 AI 编程助手直接“理解”并操作 Unity 项目，从而自动化完成场景构建、逻辑实现及测试任务。\n\n这款工具特别适合 Unity 游戏开发者、技术美术以及希望提升编码效率的独立创作者。其核心亮点在于极高的兼容性与灵活性：既支持在 Unity 编辑器中运行，也适用于运行时环境；同时完全免费且开放，无缝对接各类大模型客户端。无论是资深程序员还是正在学习游戏开发的新手，都能借助 Unity-MCP 让 AI 成为得力的结对编程伙伴，专注于创意实现而非繁琐的代码细节。","\u003Cdiv align=\"center\" width=\"100%\">\n  \u003Ch1>✨ AI Game Developer — \u003Ci>Unity SKILLS, MCP\u003C\u002Fi>\u003C\u002Fh1>\n\n[![MCP](https:\u002F\u002Fbadge.mcpx.dev 'MCP Server')](https:\u002F\u002Fmodelcontextprotocol.io\u002Fintroduction)\n[![OpenUPM](https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002Fcom.ivanmurzak.unity.mcp?label=OpenUPM&registry_uri=https:\u002F\u002Fpackage.openupm.com&labelColor=333A41 'OpenUPM package')](https:\u002F\u002Fopenupm.com\u002Fpackages\u002Fcom.ivanmurzak.unity.mcp\u002F)\n[![Docker Image](https:\u002F\u002Fimg.shields.io\u002Fdocker\u002Fimage-size\u002Fivanmurzakdev\u002Funity-mcp-server\u002Flatest?label=Docker%20Image&logo=docker&labelColor=333A41 'Docker Image')](https:\u002F\u002Fhub.docker.com\u002Fr\u002Fivanmurzakdev\u002Funity-mcp-server)\n[![Unity Editor](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FEditor-X?style=flat&logo=unity&labelColor=333A41&color=2A2A2A 'Unity Editor supported')](https:\u002F\u002Funity.com\u002Freleases\u002Feditor\u002Farchive)\n[![Unity Runtime](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FRuntime-X?style=flat&logo=unity&labelColor=333A41&color=2A2A2A 'Unity Runtime supported')](https:\u002F\u002Funity.com\u002Freleases\u002Feditor\u002Farchive)\n[![r](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fworkflows\u002Frelease\u002Fbadge.svg 'Tests Passed')](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Factions\u002Fworkflows\u002Frelease.yml)\u003C\u002Fbr>\n[![Discord](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDiscord-Join-7289da?logo=discord&logoColor=white&labelColor=333A41 'Join')](https:\u002F\u002Fdiscord.gg\u002FcfbdMZX99G)\n[![OpenUPM](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdynamic\u002Fjson?labelColor=333A41&label=Downloads&query=%24.downloads&suffix=%2Fmonth&url=https%3A%2F%2Fpackage.openupm.com%2Fdownloads%2Fpoint%2Flast-month%2Fcom.ivanmurzak.unity.mcp)](https:\u002F\u002Fopenupm.com\u002Fpackages\u002Fcom.ivanmurzak.unity.mcp\u002F)\n[![Stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FIvanMurzak\u002FUnity-MCP 'Stars')](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fstargazers)\n[![License](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002FIvanMurzak\u002FUnity-MCP?label=License&labelColor=333A41)](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002FLICENSE)\n[![Stand With Ukraine](https:\u002F\u002Fraw.githubusercontent.com\u002Fvshymanskyy\u002FStandWithUkraine\u002Fmain\u002Fbadges\u002FStandWithUkraine.svg)](https:\u002F\u002Fstand-with-ukraine.pp.ua)\n\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_f757dc570964.jpg\" alt=\"AI work\" title=\"Level building\" width=\"100%\">\n\n  \u003Cp>\n    \u003Ca href=\"https:\u002F\u002Fclaude.ai\u002Fdownload\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_c1628fa8b570.png\" alt=\"Claude\" title=\"Claude\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fopenai.com\u002Findex\u002Fintroducing-codex\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_3a15cf08fa6b.png\" alt=\"Codex\" title=\"Codex\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fwww.cursor.com\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_0b2e0b7c26c0.png\" alt=\"Cursor\" title=\"Cursor\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fcopilot\u002Foverview\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_628daf7c3c20.png\" alt=\"GitHub Copilot\" title=\"GitHub Copilot\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fgemini.google.com\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_19ba6dae62cf.png\" alt=\"Gemini\" title=\"Gemini\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fantigravity.google\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_16b7b5ad265e.png\" alt=\"Antigravity\" title=\"Antigravity\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fcode.visualstudio.com\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_9e4050cba6c1.png\" alt=\"VS Code\" title=\"VS Code\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fwww.jetbrains.com\u002Frider\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_b07d384fb907.png\" alt=\"Rider\" title=\"Rider\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fvisualstudio.microsoft.com\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_c8bfd5ccde25.png\" alt=\"Visual Studio\" title=\"Visual Studio\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-code\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_551638d72329.png\" alt=\"Open Code\" title=\"Open Code\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcline\u002Fcline\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_3ae13d2ce719.png\" alt=\"Cline\" title=\"Cline\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FKilo-Org\u002Fkilocode\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_eebfc9aa2aac.png\" alt=\"Kilo Code\" title=\"Kilo Code\" height=\"36\">\u003C\u002Fa>\n  \u003C\u002Fp>\n\n  \u003Cb>[中文](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002FREADME.zh-CN.md) | [日本語](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002FREADME.ja.md) | [Español](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002FREADME.es.md)\u003C\u002Fb>\n\n\u003C\u002Fdiv>\n\n`Unity MCP` is an AI-powered game development assistant **for Editor & Runtime**. Connect **Claude**, **Cursor**, & **Windsurf** to Unity via MCP. Automate workflows, generate code, and **enable AI within your games**.\n\nUnlike other tools, this plugin works **inside your compiled game**, allowing for real-time AI debugging and player-AI interaction.\n\n> **[💬 Join our Discord Server](https:\u002F\u002Fdiscord.gg\u002FcfbdMZX99G)** - Ask questions, showcase your work, and connect with other developers!\n\n## ![AI Game Developer — Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-features.svg?raw=true)\n\n- ✔️ **AI agents** - Use the best agents from **Anthropic**, **OpenAI**, **Microsoft**, or any other provider with no vendor lock-in\n- ✔️ **Tools** - A wide range of default [MCP Tools](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fdefault-mcp-tools.md) for operating in Unity Editor\n- ✔️ **Skills** - Generate skills for AI based on operating system, Unity version, plugins in the project\n- ✔️ **Code and Tests** - Develop game mechanics and test them with AI agents\n- ✔️ **Runtime (in-game)** - Use LLMs directly inside your compiled game for dynamic NPC behavior or debugging\n- ✔️ **Debug support** - Let AI debug and fix the problems in a project\n- ✔️ **Natural conversation** - Chat with AI like you would with a human\n- ✔️ **Flexible deployment** - Works locally (stdio) and remotely (http) via configuration\n- ✔️ **Extensible** - [Create custom Tools in your project code](#add-custom-tool)\n\n[![DOWNLOAD INSTALLER](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fbutton\u002Fbutton_download.svg?raw=true)](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Freleases\u002Flatest\u002Fdownload\u002FAI-Game-Dev-Installer.unitypackage)\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F228baf4d-4f00-4dce-939d-fb985ebdd8dd\n\n### OR use cli\n\n```bash\n# 1. Install unity-mcp-cli\nnpm install -g unity-mcp-cli\n\n# 2. Install \"AI Game Developer\" in Unity project\nunity-mcp-cli install-plugin .\u002FMyUnityProject\n\n# 3. Login to cloud server\nunity-mcp-cli login .\u002FMyUnityProject\n\n# 4. Open Unity project (auto-connects and generates skills)\nunity-mcp-cli open .\u002FMyUnityProject\n```\n\n![AI Game Developer Windows](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_8cea62cbb3eb.png)\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# Quick Start\n\nGet up and running in three steps:\n\n1. **[Install plugin](#step-1-install-unity-mcp-plugin)** — download the `.unitypackage` installer or run `openupm add com.ivanmurzak.unity.mcp`\n   > **Alternative:** `npx unity-mcp-cli install-plugin .\u002FMyUnityProject` — see [CLI documentation](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md)\n2. **[Pick an AI agent](#step-2-install-ai-agent)** — Claude Code, Claude Desktop, GitHub Copilot, Cursor, or any other\n3. **[Setup AI agent](#step-3-configure-ai-agent)** — open `Window\u002FAI Game Developer` in Unity and click **Auto-generate skills** (recommended) or **Configure MCP**\n  ![Setup AI Skills](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_33132d322c64.jpg)\n   > **Alternative:** `npx unity-mcp-cli setup-skills claude-code .\u002FMyUnityProject` — see [CLI documentation](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md)\n\n\nThat's it. Ask your AI *\"Create 3 cubes in a circle with radius 2\"* and watch it happen. ✨\n\n---\n\n# Skills and Tools Reference\n\nThe plugin ships with 100+ built-in tools across three categories. Each tool brings AI skill. All tools are available immediately after installation — no extra configuration required. See [docs\u002Fdefault-mcp-tools.md](docs\u002Fdefault-mcp-tools.md) for the full reference with detailed descriptions.\n\n\u003Cdetails>\n  \u003Csummary>Project & Assets\u003C\u002Fsummary>\n\n- `assets-copy` - Copy the asset at path and stores it at newPath\n- `assets-create-folder` - Creates a new folder in the specified parent folder\n- `assets-delete` - Delete the assets at paths from the project\n- `assets-find` - Search the asset database using the search filter string\n- `assets-find-built-in` - Search the built-in assets of the Unity Editor\n- `assets-get-data` - Get asset data from the asset file including all serializable fields and properties\n- `assets-material-create` - Create new material asset with default parameters\n- `assets-modify` - Modify asset file in the project\n- `assets-move` - Move the assets at paths in the project (also used for rename)\n- `assets-prefab-close` - Close currently opened prefab\n- `assets-prefab-create` - Create a prefab from a GameObject in the current active scene\n- `assets-prefab-instantiate` - Instantiates prefab in the current active scene\n- `assets-prefab-open` - Open prefab edit mode for a specific GameObject\n- `assets-prefab-save` - Save a prefab in prefab editing mode\n- `assets-refresh` - Refreshes the AssetDatabase\n- `assets-shader-list-all` - List all available shaders in the project assets and packages\n- `package-add` - Install a package from the Unity Package Manager registry, Git URL, or local path\n- `package-list` - List all packages installed in the Unity project (UPM packages)\n- `package-remove` - Remove (uninstall) a package from the Unity project\n- `package-search` - Search for packages in both Unity Package Manager registry and installed packages\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>Scene & Hierarchy\u003C\u002Fsummary>\n\n- `gameobject-component-add` - Add Component to GameObject\n- `gameobject-component-destroy` - Destroy one or many components from target GameObject\n- `gameobject-component-get` - Get detailed information about a specific Component on a GameObject\n- `gameobject-component-list-all` - List C# class names extended from UnityEngine.Component\n- `gameobject-component-modify` - Modify a specific Component on a GameObject\n- `gameobject-create` - Create a new GameObject in opened Prefab or in a Scene\n- `gameobject-destroy` - Destroy GameObject and all nested GameObjects recursively\n- `gameobject-duplicate` - Duplicate GameObjects in opened Prefab or in a Scene\n- `gameobject-find` - Finds specific GameObject by provided information\n- `gameobject-modify` - Modify GameObjects and\u002For attached component's fields and properties\n- `gameobject-set-parent` - Set parent GameObject to list of GameObjects\n- `object-get-data` - Get data of the specified Unity Object\n- `object-modify` - Modify the specified Unity Object\n- `scene-create` - Create new scene in the project assets\n- `scene-get-data` - Retrieves the list of root GameObjects in the specified scene\n- `scene-list-opened` - Returns the list of currently opened scenes in Unity Editor\n- `scene-open` - Open scene from the project asset file\n- `scene-save` - Save opened scene to the asset file\n- `scene-set-active` - Set the specified opened scene as the active scene\n- `scene-unload` - Unload scene from the opened scenes in Unity Editor\n- `screenshot-camera` - Captures a screenshot from a camera and returns it as an image\n- `screenshot-game-view` - Captures a screenshot from the Unity Editor Game View\n- `screenshot-scene-view` - Captures a screenshot from the Unity Editor Scene View\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>Scripting & Editor\u003C\u002Fsummary>\n\n- `console-get-logs` - Retrieves Unity Editor logs with filtering options\n- `editor-application-get-state` - Returns information about the Unity Editor application state (playmode, paused, compilation)\n- `editor-application-set-state` - Control the Unity Editor application state (start\u002Fstop\u002Fpause playmode)\n- `editor-selection-get` - Get information about the current Selection in the Unity Editor\n- `editor-selection-set` - Set the current Selection in the Unity Editor\n- `reflection-method-call` - Call any C# method with input parameters and return results\n- `reflection-method-find` - Find method in the project using C# Reflection (even private methods)\n- `script-delete` - Delete the script file(s)\n- `script-execute` - Compiles and executes C# code dynamically using Roslyn\n- `script-read` - Reads the content of a script file\n- `script-update-or-create` - Updates or creates script file with the provided C# code\n- `tests-run` - Execute Unity tests (EditMode\u002FPlayMode) with filtering and detailed results\n\n\u003C\u002Fdetails>\n\n## Install Additional Skills and Tools\n\nInstall extensions when need more tools or [create your own tools](#add-custom-tool).\n\n| Extension | Description |\n| --- | --- |\n| **[AI Animation](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-AI-Animation\u002F)** | Set of additional tools for Unity Animations |\n| **[AI ParticleSystem](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-AI-ParticleSystem\u002F)** | Set of additional tools for Unity Particle System |\n| **[AI ProBuilder](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-AI-ProBuilder\u002F)** | Set of additional tools for Unity ProBuilder |\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# Contents\n\n- [Quick Start](#quick-start)\n- [Skills and Tools Reference](#skills-and-tools-reference)\n  - [Install Additional Skills and Tools](#install-additional-skills-and-tools)\n- [Contents](#contents)\n  - [More Documentation](#more-documentation)\n- [Installation](#installation)\n  - [Step 1: Install `Unity MCP Plugin`](#step-1-install-unity-mcp-plugin)\n    - [Option 1 - Installer](#option-1---installer)\n    - [Option 2 - CLI (recommended)](#option-2---cli-recommended)\n  - [Step 2: Install `AI agent`](#step-2-install-ai-agent)\n  - [Step 3: Configure `AI agent`](#step-3-configure-ai-agent)\n    - [Automatic configuration](#automatic-configuration)\n    - [Manual configuration](#manual-configuration)\n      - [Command line configuration](#command-line-configuration)\n- [AI Workflow Examples](#ai-workflow-examples)\n  - [Advanced Features for LLM](#advanced-features-for-llm)\n    - [Core Capabilities](#core-capabilities)\n    - [Reflection-Powered Features](#reflection-powered-features)\n- [Customize Tools](#customize-tools)\n  - [Add custom `Tool`](#add-custom-tool)\n  - [Add custom `MCP Prompt`](#add-custom-mcp-prompt)\n- [Runtime usage (in-game)](#runtime-usage-in-game)\n  - [Sample: AI powered Chess game bot](#sample-ai-powered-chess-game-bot)\n  - [Why runtime usage is needed?](#why-runtime-usage-is-needed)\n- [Unity `MCP Server` setup](#unity-mcp-server-setup)\n  - [Variables](#variables)\n  - [Plugin Variables](#plugin-variables)\n  - [Docker 📦](#docker-)\n    - [`streamableHttp` Transport](#streamablehttp-transport)\n    - [`stdio` Transport](#stdio-transport)\n    - [Custom `port`](#custom-port)\n  - [Binary executable](#binary-executable)\n- [How Unity MCP Architecture Works](#how-unity-mcp-architecture-works)\n  - [What is `MCP`](#what-is-mcp)\n  - [What is `AI agent`](#what-is-ai-agent)\n  - [What is `MCP Server`](#what-is-mcp-server)\n  - [What is `MCP Tool`](#what-is-mcp-tool)\n    - [When to use `Tool`](#when-to-use-tool)\n  - [What is `MCP Resource`](#what-is-mcp-resource)\n    - [When to use `MCP Resource`](#when-to-use-mcp-resource)\n  - [What is `MCP Prompt`](#what-is-mcp-prompt)\n    - [When to use `MCP Prompt`](#when-to-use-mcp-prompt)\n- [Contribution 💙💛](#contribution-)\n\n## More Documentation\n\n| Document | Description |\n| -------- | ----------- |\n| [Default MCP Tools](docs\u002Fdefault-mcp-tools.md) | Full reference of all built-in tools with descriptions |\n| [MCP Server Setup](docs\u002Fmcp-server.md) | Server configuration, environment variables, remote hosting |\n| [Docker Deployment](docs\u002FDOCKER_DEPLOYMENT.md) | Step-by-step Docker deployment guide |\n| [Development Guide](docs\u002Fdev\u002FDevelopment.md) | Architecture, code style, CI\u002FCD — for contributors |\n| [Wiki](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fwiki) | Getting started, tutorials, API reference, FAQ |\n| [CLI Tool](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md) | Install plugins, configure, and connect via command line |\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# Installation\n\n## Step 1: Install `Unity MCP Plugin`\n\n\u003Cdetails>\n  \u003Csummary>\u003Cb>⚠️ Requirements (click)\u003C\u002Fb>\u003C\u002Fsummary>\n\n> [!IMPORTANT]\n> **Project path cannot contain spaces**\n>\n> - ✅ `C:\u002FMyProjects\u002FMyProject`\n> - ❌ `C:\u002FMy Projects\u002FMyProject`\n> - ❌ `C:\u002FMy Projects\u002FMy Project`\n> - ❌ `C:\u002FMyProjects\u002FMy Project`\n\n\u003C\u002Fdetails>\n\n### Option 1 - Installer\n\n- **[⬇️ Download Installer](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Freleases\u002Flatest\u002Fdownload\u002FAI-Game-Dev-Installer.unitypackage)**\n- **📂 Import installer into Unity project**\n  > - You can double-click on the file - Unity will open it automatically\n  > - OR: Open Unity Editor first, then click on `Assets\u002FImport Package\u002FCustom Package`, and choose the file\n\n### Option 2 - CLI (recommended)\n\nInstall the plugin via [`unity-mcp-cli`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md) — no Unity Editor needed:\n\n```bash\n# 1.1 Install unity-mcp-cli                                #  ┌────────────────────┐\nnpm install -g unity-mcp-cli                               #  │ Available AI agent │\n                                                           #  ├────────────────────┤\n# 1.2 (Optional) Install Unity                             #  │ antigravity        │\nunity-mcp-cli install-unity                                #  │ claude-code        │\n                                                           #  │ claude-desktop     │\n# 1.3 (Optional) Create Unity project                      #  │ cline              │\nunity-mcp-cli create-project .\u002FMyUnityProject              #  │ codex              │\n                                                           #  │ cursor             │\n# 2. Install \"AI Game Developer\" in Unity project          #  │ gemini             │\nunity-mcp-cli install-plugin .\u002FMyUnityProject              #  │ github-copilot-cli │\n                                                           #  │ kilo-code          │\n# 3. Login to cloud server                                 #  │ open-code          │\nunity-mcp-cli login .\u002FMyUnityProject                       #  │ rider-junie        │\n                                                           #  │ unity-ai           │\n# 4. Open Unity project (auto-connects and generates skills)  │ vs-copilot         │\nunity-mcp-cli open .\u002FMyUnityProject                        #  │ vscode-copilot     │\n                                                           #  └────────────────────┘\n# 5. Wait for Unity Editor to be ready\nunity-mcp-cli wait-for-ready .\u002FMyUnityProject\n```\n\n> See [full CLI documentation](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md) for all available commands.\n\n## Step 2: Install `AI agent`\n\nChoose a single `AI agent` you prefer - you don't need to install all of them. This will be your main chat window to communicate with the LLM.\n\n- [Claude Code](https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-code) **(recommended)**\n- [Claude Desktop](https:\u002F\u002Fclaude.ai\u002Fdownload)\n- [GitHub Copilot in VS Code](https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fcopilot\u002Foverview)\n- [Antigravity](https:\u002F\u002Fantigravity.google\u002F)\n- [Cursor](https:\u002F\u002Fwww.cursor.com\u002F)\n- [Windsurf](https:\u002F\u002Fwindsurf.com)\n- Any other supported\n\n> The AI Game Developer is quite universal, which is why you may use any AI agent you prefer - it will work as smoothly as any other. The only important requirement is that the AI agent must support Skills or dynamic MCP Tool updates.\n\n## Step 3: Configure `AI agent`\n\n### Automatic configuration\n\n- Open Unity project\n- Open `Window\u002FAI Game Developer`\n- Option 1: Click `Auto-generate` Skills **(recommended)**\n- Option 2: Click `Configure` Model Context Protocol (MCP)\n\n![Unity_AI](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_0efd8c6c3cea.gif)\n\n> If your MCP client is not in the list, use the raw JSON shown in the window to inject it into your MCP client. Read the instructions for your specific MCP client on how to do this.\n\n### Manual configuration\n\nIf automatic configuration doesn't work for you for any reason, use the JSON from the `AI Game Developer (Unity-MCP)` window to configure any `MCP Client` manually.\n\n#### Command line configuration\n\n\u003Cdetails>\n  \u003Csummary>\u003Cb>Create \u003Ccode>command\u003C\u002Fcode>\u003C\u002Fb>\u003C\u002Fsummary>\n\n**1. Choose your `\u003Ccommand>` for your environment**\n\n| Platform            | `\u003Ccommand>`                                                                                                 |\n| ------------------- | ----------------------------------------------------------------------------------------------------------- |\n| Windows x64         | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fwin-x64\u002Funity-mcp-server.exe\" port=\u003Cport> client-transport=stdio`   |\n| Windows x86         | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fwin-x86\u002Funity-mcp-server.exe\" port=\u003Cport> client-transport=stdio`   |\n| Windows arm64       | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fwin-arm64\u002Funity-mcp-server.exe\" port=\u003Cport> client-transport=stdio` |\n| MacOS Apple-Silicon | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fosx-arm64\u002Funity-mcp-server\" port=\u003Cport> client-transport=stdio`     |\n| MacOS Apple-Intel   | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fosx-x64\u002Funity-mcp-server\" port=\u003Cport> client-transport=stdio`       |\n| Linux x64           | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Flinux-x64\u002Funity-mcp-server\" port=\u003Cport> client-transport=stdio`     |\n| Linux arm64         | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Flinux-arm64\u002Funity-mcp-server\" port=\u003Cport> client-transport=stdio`   |\n\n**2. Replace `\u003CunityProjectPath>` with the full path to Unity project**\n\n**3. Replace `\u003Cport>` with your port from AI Game Developer configuration**\n\n**4. Add MCP server using command line**\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_19ba6dae62cf.png\" width=\"16\" height=\"16\" alt=\"Gemini CLI\"> Gemini CLI\u003C\u002Fsummary>\n\n  ```bash\n  gemini mcp add ai-game-developer \u003Ccommand>\n  ```\n\n  > Replace `\u003Ccommand>` from the table above\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_c1628fa8b570.png\" width=\"16\" height=\"16\" alt=\"Claude Code CLI\"> Claude Code CLI\u003C\u002Fsummary>\n\n  ```bash\n  claude mcp add ai-game-developer \u003Ccommand>\n  ```\n\n  > Replace `\u003Ccommand>` from the table above\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_628daf7c3c20.png\" width=\"16\" height=\"16\" alt=\"GitHub Copilot CLI\"> GitHub Copilot CLI\u003C\u002Fsummary>\n\n  ```bash\n  copilot\n  ```\n\n  ```bash\n  \u002Fmcp add\n  ```\n\n  Server name: `ai-game-developer`\n  Server type: `local`\n  Command: `\u003Ccommand>`\n  > Replace `\u003Ccommand>` from the table above\n\u003C\u002Fdetails>\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# AI Workflow Examples\n\nCommunicate with the AI (LLM) in your `AI agent`. Ask it to do anything you want. The better you describe your task or idea, the better it will perform the job.\n\nSome `AI agents` allow you to choose different LLM models. Pay attention to this feature, as some models may work much better than others.\n\n**Example prompts:**\n\n```text\nExplain my scene hierarchy\n```\n\n```text\nCreate 3 spheres on top of each other\n```\n\n```text\nCreate metallic golden material and attach it to a new sphere gameObject\n```\n\n> Make sure `Agent` mode is enabled if using VS Code with Copilot\n\n## Advanced Features for LLM\n\nUnity MCP provides advanced tools that enable the LLM to work faster and more effectively, avoiding mistakes and self-correcting when errors occur. Everything is designed to achieve your goals efficiently.\n\n### Core Capabilities\n\n- ✔️ **Agent-ready tools** - Find anything you need in 1-2 steps\n- ✔️ **Instant compilation** - C# code compilation & execution using `Roslyn` for faster iteration\n- ✔️ **Full asset access** - Read\u002Fwrite access to assets and C# scripts\n- ✔️ **Intelligent feedback** - Well-described positive and negative feedback for proper issue understanding\n\n### Reflection-Powered Features\n\n- ✔️ **Object references** - Provide references to existing objects for instant C# code\n- ✔️ **Project data access** - Get full access to entire project data in a readable format\n- ✔️ **Granular modifications** - Populate & modify any piece of data in the project\n- ✔️ **Method discovery** - Find any method in the entire codebase, including compiled DLL files\n- ✔️ **Method execution** - Call any method in the entire codebase\n- ✔️ **Advanced parameters** - Provide any property for method calls, even references to existing objects in memory\n- ✔️ **Live Unity API** - Unity API instantly available - even when Unity changes, you get the fresh API\n- ✔️ **Self-documenting** - Access human-readable descriptions of any `class`, `method`, or `property` via `Description` attributes\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# Customize Tools\n\n**[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)** supports custom `MCP Tool`, `MCP Resource`, and `MCP Prompt` development by project owners. The MCP server takes data from the `Unity MCP Plugin` and exposes it to a client. Anyone in the MCP communication chain will receive information about new MCP features, which the LLM may decide to use at some point.\n\n## Add custom `Tool`\n\nTo add a custom `Tool`, you need:\n\n1. A class with the `McpPluginToolType` attribute\n2. A method in the class with the `McpPluginTool` attribute\n3. *Optional:* Add a `Description` attribute to each method argument to help the LLM understand it\n4. *Optional:* Use `string? optional = null` properties with `?` and default values to mark them as `optional` for the LLM\n\n> Note that the line `MainThread.Instance.Run(() =>` allows you to run code on the main thread, which is required for interacting with Unity's API. If you don't need this and running the tool in a background thread is acceptable, avoid using the main thread for efficiency purposes.\n\n```csharp\n[McpPluginToolType]\npublic class Tool_GameObject\n{\n    [McpPluginTool\n    (\n        \"MyCustomTask\",\n        Title = \"Create a new GameObject\"\n    )]\n    [Description(\"Explain here to LLM what is this, when it should be called.\")]\n    public string CustomTask\n    (\n        [Description(\"Explain to LLM what is this.\")]\n        string inputData\n    )\n    {\n        \u002F\u002F do anything in background thread\n\n        return MainThread.Instance.Run(() =>\n        {\n            \u002F\u002F do something in main thread if needed\n\n            return $\"[Success] Operation completed.\";\n        });\n    }\n}\n```\n\n## Add custom `MCP Prompt`\n\n`MCP Prompt` allows you to inject custom prompts into the conversation with the LLM. It supports two sender roles: User and Assistant. This is a quick way to instruct the LLM to perform specific tasks. You can generate prompts using custom data, providing lists or any other relevant information.\n\n```csharp\n[McpPluginPromptType]\npublic static class Prompt_ScriptingCode\n{\n    [McpPluginPrompt(Name = \"add-event-system\", Role = Role.User)]\n    [Description(\"Implement UnityEvent-based communication system between GameObjects.\")]\n    public string AddEventSystem()\n    {\n        return \"Create event system using UnityEvents, UnityActions, or custom event delegates for decoupled communication between game systems and components.\";\n    }\n}\n```\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# Runtime usage (in-game)\n\nUse **[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)** in your game\u002Fapp. Use Tools, Resources or Prompts. By default there are no tools, you would need to implement your custom.\n\n```csharp\n\u002F\u002F Build MCP plugin\nvar mcpPlugin = UnityMcpPluginRuntime.Initialize(builder =>\n    {\n        builder.WithConfig(config =>\n        {\n            config.Host = \"http:\u002F\u002Flocalhost:8080\";\n            config.Token = \"your-token\";\n        });\n        \u002F\u002F Automatically register all tools from the current assembly\n        builder.WithToolsFromAssembly(Assembly.GetExecutingAssembly());\n    })\n    .Build();\n\nawait mcpPlugin.Connect(); \u002F\u002F Start active connection with retry to Unity-MCP-Server\n\nawait mcpPlugin.Disconnect(); \u002F\u002F Stop active connection and close existed connection\n```\n\n## Sample: AI powered Chess game bot\n\nThere is a classic Chess game. Lets outsource to LLM the bot logic. Bot should do the turn using game rules.\n\n```csharp\n[McpPluginToolType]\npublic static class ChessGameAI\n{\n    [McpPluginTool(\"chess-do-turn\", Title = \"Do the turn\")]\n    [Description(\"Do the turn in the chess game. Returns true if the turn was accepted, false otherwise.\")]\n    public static Task\u003Cbool> DoTurn(int figureId, Vector2Int position)\n    {\n        return MainThread.Instance.RunAsync(() => ChessGameController.Instance.DoTurn(figureId, position));\n    }\n\n    [McpPluginTool(\"chess-get-board\", Title = \"Get the board\")]\n    [Description(\"Get the current state of the chess board.\")]\n    public static Task\u003CBoardData> GetBoard()\n    {\n        return MainThread.Instance.RunAsync(() => ChessGameController.Instance.GetBoardData());\n    }\n}\n```\n\n## Why runtime usage is needed?\n\nThere are many use cases, lets imagine you are working on a Chess game with bot. You may outsource the bot decision making to LLM by writing few lines of code.\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# Unity `MCP Server` setup\n\n**[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)** Server supports many different launch options and Docker deployment. Both transport protocols are supported: `streamableHttp` and `stdio`. If you need to customize or deploy Unity MCP Server to a cloud, this section is for you. [Read more...](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fmcp-server.md)\n\n## Variables\n\nDoesn't matter what launch option you choose, all of them support custom configuration using both Environment Variables and Command Line Arguments. It would work with default values, if you just need to launch it, don't waste your time for the variables. Just make sure Unity Plugin also has default values, especially the `--port`, they should be equal.\n\n| Environment Variable         | Command Line Args    | Description                                                                  |\n| ---------------------------- | -------------------- | ---------------------------------------------------------------------------- |\n| `MCP_PLUGIN_PORT`            | `--port`             | **Client** -> **Server** \u003C- **Plugin** connection port (default: 8080)       |\n| `MCP_PLUGIN_CLIENT_TIMEOUT`   | `--plugin-timeout`   | **Plugin** -> **Server** connection timeout (ms) (default: 10000)            |\n| `MCP_PLUGIN_CLIENT_TRANSPORT` | `--client-transport` | **Client** -> **Server** transport type: `stdio` or `streamableHttp` (default: `streamableHttp`) |\n\n> Command line args support also the option with a single `-` prefix (`-port`) and an option without prefix at all (`port`).\n\n> **Choosing a transport:** Use `stdio` when the MCP client launches the server binary directly (local use — this is the most common setup). Use `streamableHttp` when running the server as a standalone process or in Docker\u002Fcloud, and connecting over HTTP.\n\n## Plugin Variables\n\nThe Unity MCP Plugin reads the following environment variables (and command-line arguments) on startup to override values from the saved config file. Overrides are applied at runtime; on first run or when a new authentication token is generated, the overridden values are **written to the config file**. On subsequent runs, overrides are applied in memory but are not automatically saved. The exception is `UNITY_MCP_TOOLS`, which uses `[JsonIgnore]` and is **never persisted** — it is runtime-only.\n\n| Environment Variable        | Command Line Arg            | Values              | Description                                   |\n| --------------------------- | --------------------------- | ------------------- | --------------------------------------------- |\n| `UNITY_MCP_HOST`            | `-UNITY_MCP_HOST`           | URL string                    | Override the MCP Server host URL                                    |\n| `UNITY_MCP_KEEP_CONNECTED`  | `-UNITY_MCP_KEEP_CONNECTED` | `true` \u002F `false`              | Force enable or disable the active connection                       |\n| `UNITY_MCP_AUTH_OPTION`     | `-UNITY_MCP_AUTH_OPTION`    | `none` \u002F `required`           | Force set the authentication mode                                   |\n| `UNITY_MCP_TOKEN`           | `-UNITY_MCP_TOKEN`          | string                        | Force set the authentication token                                  |\n| `UNITY_MCP_TOOLS`           | `-UNITY_MCP_TOOLS`          | comma-separated tool IDs      | Enable only the listed tools; all others are disabled. Unknown IDs are logged as errors. |\n\n> Command-line args take precedence over environment variables. Both override the saved config file value.\n\n**Example (CI\u002FCD batch mode):**\n\n```bash\nUnity.exe -batchmode -nographics \\\n  -UNITY_MCP_HOST=http:\u002F\u002Flocalhost:8080 \\\n  -UNITY_MCP_KEEP_CONNECTED=true \\\n  -UNITY_MCP_AUTH_OPTION=required \\\n  -UNITY_MCP_TOKEN=my-secret-token\n```\n\n## Docker 📦\n\n[![Docker Image](https:\u002F\u002Fimg.shields.io\u002Fdocker\u002Fimage-size\u002Fivanmurzakdev\u002Funity-mcp-server\u002Flatest?label=Docker%20Image&logo=docker&labelColor=333A41 'Docker Image')](https:\u002F\u002Fhub.docker.com\u002Fr\u002Fivanmurzakdev\u002Funity-mcp-server)\n\nMake sure Docker is installed. And please make sure Docker Desktop is launched if you are at Windows operation system.\n\n[Read advanced Docker configuration instructions](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002FDOCKER_DEPLOYMENT.md).\n\n### `streamableHttp` Transport\n\n```bash\ndocker run -p 8080:8080 ivanmurzakdev\u002Funity-mcp-server\n```\n\n\u003Cdetails>\n  \u003Csummary>\u003Ccode>MCP Client\u003C\u002Fcode> config:\u003C\u002Fsummary>\n\n```json\n{\n  \"mcpServers\": {\n    \"ai-game-developer\": {\n      \"url\": \"http:\u002F\u002Flocalhost:8080\"\n    }\n  }\n}\n```\n\n> Replace `url` with your real endpoint if it is hosted in cloud.\n\n\u003C\u002Fdetails>\n\n### `stdio` Transport\n\nFor using this variant, `MCP Client` should launch the `MCP Server` in the docker. It is achievable through the modified `MCP Client` configuration.\n\n```bash\ndocker run -t -e MCP_PLUGIN_CLIENT_TRANSPORT=stdio -p 8080:8080 ivanmurzakdev\u002Funity-mcp-server\n```\n\n\u003Cdetails>\n  \u003Csummary>\u003Ccode>MCP Client\u003C\u002Fcode> config:\u003C\u002Fsummary>\n\n```json\n{\n  \"mcpServers\": {\n    \"ai-game-developer\": {\n      \"command\": \"docker\",\n      \"args\": [\n        \"run\",\n        \"-t\",\n        \"-e\",\n        \"MCP_PLUGIN_CLIENT_TRANSPORT=stdio\",\n        \"-p\",\n        \"8080:8080\",\n        \"ivanmurzakdev\u002Funity-mcp-server\"\n      ]\n    }\n  }\n}\n```\n\n\u003C\u002Fdetails>\n\n### Custom `port`\n\n```bash\ndocker run -e MCP_PLUGIN_PORT=123 -p 123:123 ivanmurzakdev\u002Funity-mcp-server\n```\n\n\u003Cdetails>\n  \u003Csummary>\u003Ccode>MCP Client\u003C\u002Fcode> config:\u003C\u002Fsummary>\n\n```json\n{\n  \"mcpServers\": {\n    \"ai-game-developer\": {\n      \"url\": \"http:\u002F\u002Flocalhost:123\"\n    }\n  }\n}\n```\n\n> Replace `url` with your real endpoint if it is hosted in cloud\n\u003C\u002Fdetails>\n\n## Binary executable\n\nYou may launch Unity `MCP Server` directly from a binary file. You would need to have a binary compiled specifically for your CPU architecture. Check [GitHub Release Page](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Freleases), it contains pre-compiled binaries for all CPU architectures.\n\n```bash\n.\u002Funity-mcp-server --port 8080 --plugin-timeout 10000 --client-transport stdio\n```\n\n\u003Cdetails>\n  \u003Csummary>\u003Ccode>MCP Client\u003C\u002Fcode> config:\u003C\u002Fsummary>\n\n> Replace `\u003Cproject>` with your Unity project path.\n\n```json\n{\n  \"mcpServers\": {\n    \"ai-game-developer\": {\n      \"command\": \"\u003Cproject>\u002FLibrary\u002Fmcp-server\u002Fwin-x64\u002Funity-mcp-server.exe\",\n      \"args\": [\n        \"--port=8080\",\n        \"--plugin-timeout=10000\",\n        \"--client-transport=stdio\"\n      ]\n    }\n  }\n}\n```\n\n\u003C\u002Fdetails>\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# How Unity MCP Architecture Works\n\n**[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)** serves as a bridge between LLMs and Unity. It exposes and explains Unity's tools to the LLM, which then understands the interface and utilizes the tools according to user requests.\n\nConnect **[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)** to LLM clients such as [Claude](https:\u002F\u002Fclaude.ai\u002Fdownload) or [Cursor](https:\u002F\u002Fwww.cursor.com\u002F) using the integrated `AI Connector` window. Custom clients are also supported.\n\nThe system is highly extensible - you can define custom `MCP Tools`, `MCP Resource` or `MCP Prompt` directly in your Unity project codebase, exposing new capabilities to AI or automation clients. This makes Unity MCP a flexible foundation for building advanced workflows, rapid prototyping, and integrating AI-driven features into your development process.\n\n## What is `MCP`\n\nMCP - Model Context Protocol. In a few words, that is `USB Type-C` for AI, specifically for LLM (Large Language Model). It teaches LLM how to use external features. Such as Unity Engine in this case, or even your custom C# method in your code. [Official documentation](https:\u002F\u002Fmodelcontextprotocol.io\u002F).\n\n## What is `AI agent`\n\nIt is an application with a chat window. It may have smart agents to operate better, it may have embedded advanced MCP Tools. In general well done MCP Client is 50% of the AI success of executing a task. That is why it is very important to choose the best one for usage.\n\n## What is `MCP Server`\n\nIt is a bridge between `MCP Client` and \"something else\", in this particular case it is Unity Engine. This project includes `MCP Server`.\n\n## What is `MCP Tool`\n\n`MCP Tool` is a function or method that the LLM can call to interact with Unity. These tools act as the bridge between natural language requests and actual Unity operations. When you ask the AI to \"create a cube\" or \"change material color,\" it uses MCP Tools to execute these actions.\n\n**Key characteristics:**\n\n- **Executable functions** that perform specific operations\n- **Typed parameters** with descriptions to help the LLM understand what data to provide\n- **Return values** that give feedback about the operation's success or failure\n- **Thread-aware** - can run on main thread for Unity API calls or background thread for heavy processing\n\n### When to use `Tool`\n\n- **Automate repetitive tasks** - Create tools for common operations you do frequently\n- **Complex operations** - Bundle multiple Unity API calls into a single, easy-to-use tool\n- **Project-specific workflows** - Build tools that understand your project's specific structure and conventions\n- **Error-prone tasks** - Create tools that include validation and error handling\n- **Custom game logic** - Expose your game's systems to AI for dynamic content creation\n\n**Examples:**\n\n- Creating and configuring GameObjects with specific components\n- Batch processing assets (textures, materials, prefabs)\n- Setting up lighting and post-processing effects\n- Generating level geometry or placing objects procedurally\n- Configuring physics settings or collision layers\n\n## What is `MCP Resource`\n\n`MCP Resource` provides read-only access to data within your Unity project. Unlike MCP Tools that perform actions, Resources allow the LLM to inspect and understand your project's current state, assets, and configuration. Think of them as \"sensors\" that give the AI context about your project.\n\n**Key characteristics:**\n\n- **Read-only access** to project data and Unity objects\n- **Structured information** presented in a format the LLM can understand\n- **Real-time data** that reflects the current state of your project\n- **Contextual awareness** helping the AI make informed decisions\n\n### When to use `MCP Resource`\n\n- **Project analysis** - Let AI understand your project structure, assets, and organization\n- **Debugging assistance** - Provide current state information for troubleshooting\n- **Intelligent suggestions** - Give AI context to make better recommendations\n- **Documentation generation** - Automatically create documentation based on project state\n- **Asset management** - Help AI understand what assets are available and their properties\n\n**Examples:**\n\n- Exposing scene hierarchy and GameObject properties\n- Listing available materials, textures, and their settings\n- Showing script dependencies and component relationships\n- Displaying current lighting setup and render pipeline configuration\n- Providing information about audio sources, animations, and particle systems\n\n## What is `MCP Prompt`\n\n`MCP Prompt` allows you to inject pre-defined prompts into the conversation with the LLM. These are smart templates that can provide context, instructions, or knowledge to guide the AI's behavior. Prompts can be static text or dynamically generated based on your project's current state.\n\n**Key characteristics:**\n\n- **Contextual guidance** that influences how the AI responds\n- **Role-based** - can simulate different personas (User requests or Assistant knowledge)\n- **Dynamic content** - can include real-time project data\n- **Reusable templates** for common scenarios and workflows\n\n### When to use `MCP Prompt`\n\n- **Provide domain knowledge** - Share best practices and coding standards specific to your project\n- **Set coding conventions** - Establish naming conventions, architecture patterns, and code style\n- **Give context about project structure** - Explain how your project is organized and why\n- **Share workflow instructions** - Provide step-by-step procedures for common tasks\n- **Inject specialized knowledge** - Add information about specific Unity features, third-party assets, or custom systems\n\n**Examples:**\n\n- \"Always use PascalCase for public methods and camelCase for private fields\"\n- \"This project uses a custom event system located in Scripts\u002FEvents\u002F\"\n- \"When creating UI elements, always add them to the Canvas in Scene\u002FUI\u002FMainCanvas\"\n- \"Performance is critical - prefer object pooling for frequently instantiated objects\"\n- \"This project follows SOLID principles - explain any architecture decisions\"\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# Contribution 💙💛\n\nContributions are highly appreciated. Bring your ideas and let's make game development simpler than ever before! Do you have an idea for a new `Tool` or feature, or did you spot a bug and know how to fix it?\n\n**Please give this project a star 🌟 if you find it useful!**\n\n1. 👉 [Read Development documentation](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fdev\u002FDevelopment.md)\n2. 👉 [Fork the project](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Ffork)\n3. Clone the fork and open the `.\u002FUnity-MCP-Plugin` folder in Unity\n4. Implement new things in the project, commit, push it to GitHub\n5. Create Pull Request targeting original [Unity-MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare) repository, `main` branch.\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n","\u003Cdiv align=\"center\" width=\"100%\">\n  \u003Ch1>✨ AI 游戏开发者 — \u003Ci>Unity 技能、MCP\u003C\u002Fi>\u003C\u002Fh1>\n\n[![MCP](https:\u002F\u002Fbadge.mcpx.dev 'MCP 服务器')](https:\u002F\u002Fmodelcontextprotocol.io\u002Fintroduction)\n[![OpenUPM](https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002Fcom.ivanmurzak.unity.mcp?label=OpenUPM&registry_uri=https:\u002F\u002Fpackage.openupm.com&labelColor=333A41 'OpenUPM 包')](https:\u002F\u002Fopenupm.com\u002Fpackages\u002Fcom.ivanmurzak.unity.mcp\u002F)\n[![Docker 镜像](https:\u002F\u002Fimg.shields.io\u002Fdocker\u002Fimage-size\u002Fivanmurzakdev\u002Funity-mcp-server\u002Flatest?label=Docker%20Image&logo=docker&labelColor=333A41 'Docker 镜像')](https:\u002F\u002Fhub.docker.com\u002Fr\u002Fivanmurzakdev\u002Funity-mcp-server)\n[![Unity 编辑器](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FEditor-X?style=flat&logo=unity&labelColor=333A41&color=2A2A2A '支持 Unity 编辑器')](https:\u002F\u002Funity.com\u002Freleases\u002Feditor\u002Farchive)\n[![Unity 运行时](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FRuntime-X?style=flat&logo=unity&labelColor=333A41&color=2A2A2A '支持 Unity 运行时')](https:\u002F\u002Funity.com\u002Freleases\u002Feditor\u002Farchive)\n[![r](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fworkflows\u002Frelease\u002Fbadge.svg '测试通过')](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Factions\u002Fworkflows\u002Frelease.yml)\u003C\u002Fbr>\n[![Discord](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDiscord-Join-7289da?logo=discord&logoColor=white&labelColor=333A41 '加入')](https:\u002F\u002Fdiscord.gg\u002FcfbdMZX99G)\n[![OpenUPM](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdynamic\u002Fjson?labelColor=333A41&label=Downloads&query=%24.downloads&suffix=%2Fmonth&url=https%3A%2F%2Fpackage.openupm.com%2Fdownloads%2Fpoint%2Flast-month%2Fcom.ivanmurzak.unity.mcp)](https:\u002F\u002Fopenupm.com\u002Fpackages\u002Fcom.ivanmurzak.unity.mcp\u002F)\n[![星标](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002FIvanMurzak\u002FUnity-MCP '星标')](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fstargazers)\n[![许可证](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002FIvanMurzak\u002FUnity-MCP?label=License&labelColor=333A41)](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002FLICENSE)\n[![与乌克兰同在](https:\u002F\u002Fraw.githubusercontent.com\u002Fvshymanskyy\u002FStandWithUkraine\u002Fmain\u002Fbadges\u002FStandWithUkraine.svg)](https:\u002F\u002Fstand-with-ukraine.pp.ua)\n\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_f757dc570964.jpg\" alt=\"AI 工作\" title=\"关卡构建\" width=\"100%\">\n\n  \u003Cp>\n    \u003Ca href=\"https:\u002F\u002Fclaude.ai\u002Fdownload\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_c1628fa8b570.png\" alt=\"Claude\" title=\"Claude\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fopenai.com\u002Findex\u002Fintroducing-codex\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_3a15cf08fa6b.png\" alt=\"Codex\" title=\"Codex\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fwww.cursor.com\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_0b2e0b7c26c0.png\" alt=\"Cursor\" title=\"Cursor\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fcopilot\u002Foverview\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_628daf7c3c20.png\" alt=\"GitHub Copilot\" title=\"GitHub Copilot\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fgemini.google.com\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_19ba6dae62cf.png\" alt=\"Gemini\" title=\"Gemini\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fantigravity.google\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_16b7b5ad265e.png\" alt=\"Antigravity\" title=\"Antigravity\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fcode.visualstudio.com\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_9e4050cba6c1.png\" alt=\"VS Code\" title=\"VS Code\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fwww.jetbrains.com\u002Frider\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_b07d384fb907.png\" alt=\"Rider\" title=\"Rider\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fvisualstudio.microsoft.com\u002F\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_c8bfd5ccde25.png\" alt=\"Visual Studio\" title=\"Visual Studio\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-code\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_551638d72329.png\" alt=\"Open Code\" title=\"Open Code\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fcline\u002Fcline\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_3ae13d2ce719.png\" alt=\"Cline\" title=\"Cline\" height=\"36\">\u003C\u002Fa>  \n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FKilo-Org\u002Fkilocode\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_eebfc9aa2aac.png\" alt=\"Kilo Code\" title=\"Kilo Code\" height=\"36\">\u003C\u002Fa>\n  \u003C\u002Fp>\n\n  \u003Cb>[中文](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002FREADME.zh-CN.md) | [日本語](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002FREADME.ja.md) | [Español](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002FREADME.es.md)\u003C\u002Fb>\n\n\u003C\u002Fdiv>\n\n`Unity MCP` 是一款基于 AI 的游戏开发助手，**适用于编辑器和运行时**。通过 MCP 将 **Claude**、**Cursor** 和 **Windsurf** 连接到 Unity。自动化工作流、生成代码，并在您的游戏中**启用 AI 功能**。\n\n与其他工具不同，此插件可在**已编译的游戏内部运行**，从而实现实时的 AI 调试以及玩家与 AI 的互动。\n\n> **[💬 加入我们的 Discord 服务器](https:\u002F\u002Fdiscord.gg\u002FcfbdMZX99G)** - 提问、展示作品并与开发者交流！\n\n## ![AI 游戏开发者 — Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-features.svg?raw=true)\n\n- ✔️ **AI 代理** - 使用来自 **Anthropic**、**OpenAI**、**Microsoft** 或其他任何提供商的最佳代理，无供应商锁定\n- ✔️ **工具** - 丰富的默认 [MCP 工具](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fdefault-mcp-tools.md)，用于在 Unity 编辑器中操作\n- ✔️ **技能** - 根据操作系统、Unity 版本以及项目中的插件，为 AI 生成相应的技能\n- ✔️ **代码与测试** - 利用 AI 代理开发游戏机制并进行测试\n- ✔️ **运行时（游戏内）** - 直接在已编译的游戏内部使用 LLM，实现动态 NPC 行为或调试\n- ✔️ **调试支持** - 让 AI 调试并修复项目中的问题\n- ✔️ **自然对话** - 像与人类交谈一样与 AI 对话\n- ✔️ **灵活部署** - 可通过配置在本地（stdio）和远程（http）运行\n- ✔️ **可扩展性** - [在项目代码中创建自定义工具](#add-custom-tool)\n\n[![下载安装程序](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fbutton\u002Fbutton_download.svg?raw=true)](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Freleases\u002Flatest\u002Fdownload\u002FAI-Game-Dev-Installer.unitypackage)\n\nhttps:\u002F\u002Fgithub.com\u002Fuser-attachments\u002Fassets\u002F228baf4d-4f00-4dce-939d-fb985ebdd8dd\n\n### 或者使用命令行\n\n```bash\n# 1. 安装 unity-mcp-cli\nnpm install -g unity-mcp-cli\n\n# 2. 在 Unity 项目中安装“AI 游戏开发者”\nunity-mcp-cli install-plugin .\u002FMyUnityProject\n\n# 3. 登录云服务器\nunity-mcp-cli login .\u002FMyUnityProject\n\n# 4. 打开 Unity 项目（自动连接并生成技能）\nunity-mcp-cli open .\u002FMyUnityProject\n```\n\n![AI 游戏开发者 Windows](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_8cea62cbb3eb.png)\n\n![AI 游戏开发者 — Unity 技能与 MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# 快速入门\n\n只需三步即可快速上手：\n\n1. **[安装插件](#step-1-install-unity-mcp-plugin)** — 下载 `.unitypackage` 安装包，或运行 `openupm add com.ivanmurzak.unity.mcp`\n   > **替代方案：** `npx unity-mcp-cli install-plugin .\u002FMyUnityProject` — 请参阅 [CLI 文档](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md)\n2. **[选择 AI 助手](#step-2-install-ai-agent)** — Claude Code、Claude Desktop、GitHub Copilot、Cursor 或其他任何助手\n3. **[设置 AI 助手](#step-3-configure-ai-agent)** — 在 Unity 中打开 `Window\u002FAI 游戏开发者`，然后点击 **自动生成技能**（推荐）或 **配置 MCP**\n  ![设置 AI 技能](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_33132d322c64.jpg)\n   > **替代方案：** `npx unity-mcp-cli setup-skills claude-code .\u002FMyUnityProject` — 请参阅 [CLI 文档](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md)\n\n\n就是这样。只需让你的 AI 输入 *\"创建一个半径为 2 的圆圈，并在其中放置 3 个立方体\"*，它就会立即完成！✨\n\n---\n\n# 技能与工具参考\n\n该插件内置了三大类共 100 多种工具。每种工具都能为 AI 增添一项技能。所有工具在安装后即可立即使用，无需额外配置。完整的参考列表及详细说明，请参阅 [docs\u002Fdefault-mcp-tools.md](docs\u002Fdefault-mcp-tools.md)。\n\n\u003Cdetails>\n  \u003Csummary>项目与资源\u003C\u002Fsummary>\n\n- `assets-copy` - 复制指定路径下的资源，并将其存储到新路径\n- `assets-create-folder` - 在指定父文件夹中创建新文件夹\n- `assets-delete` - 从项目中删除指定路径下的资源\n- `assets-find` - 使用搜索过滤字符串在资源数据库中搜索资源\n- `assets-find-built-in` - 搜索 Unity 编辑器的内置资源\n- `assets-get-data` - 从资源文件中获取数据，包括所有可序列化的字段和属性\n- `assets-material-create` - 创建具有默认参数的新材质资源\n- `assets-modify` - 修改项目中的资源文件\n- `assets-move` - 在项目中移动指定路径下的资源（也可用于重命名）\n- `assets-prefab-close` - 关闭当前打开的预制件\n- `assets-prefab-create` - 从当前活动场景中的 GameObject 创建预制件\n- `assets-prefab-instantiate` - 在当前活动场景中实例化预制件\n- `assets-prefab-open` - 打开特定 GameObject 的预制件编辑模式\n- `assets-prefab-save` - 在预制件编辑模式下保存预制件\n- `assets-refresh` - 刷新 AssetDatabase\n- `assets-shader-list-all` - 列出项目资源和包中所有可用着色器\n- `package-add` - 从 Unity Package Manager 注册表、Git URL 或本地路径安装包\n- `package-list` - 列出 Unity 项目中已安装的所有包（UPM 包）\n- `package-remove` - 从 Unity 项目中移除（卸载）某个包\n- `package-search` - 在 Unity Package Manager 注册表和已安装的包中搜索包\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>场景与层级视图\u003C\u002Fsummary>\n\n- `gameobject-component-add` - 向 GameObject 添加组件\n- `gameobject-component-destroy` - 销毁目标 GameObject 上的一个或多个组件\n- `gameobject-component-get` - 获取 GameObject 上特定组件的详细信息\n- `gameobject-component-list-all` - 列出继承自 UnityEngine.Component 的 C# 类名\n- `gameobject-component-modify` - 修改 GameObject 上的特定组件\n- `gameobject-create` - 在打开的预制件或场景中创建新的 GameObject\n- `gameobject-destroy` - 销毁 GameObject 及其所有嵌套的 GameObject（递归操作）\n- `gameobject-duplicate` - 在打开的预制件或场景中复制 GameObject\n- `gameobject-find` - 根据提供的信息查找特定 GameObject\n- `gameobject-modify` - 修改 GameObject 及其附加组件的字段和属性\n- `gameobject-set-parent` - 将一组 GameObject 的父级设置为指定 GameObject\n- `object-get-data` - 获取指定 Unity 对象的数据\n- `object-modify` - 修改指定的 Unity 对象\n- `scene-create` - 在项目资源中创建新场景\n- `scene-get-data` - 获取指定场景中根 GameObject 的列表\n- `scene-list-opened` - 返回 Unity 编辑器中当前打开的场景列表\n- `scene-open` - 从项目资源文件中打开场景\n- `scene-save` - 将打开的场景保存到资源文件中\n- `scene-set-active` - 将指定的打开场景设置为活动场景\n- `scene-unload` - 从 Unity 编辑器的已打开场景中卸载场景\n- `screenshot-camera` - 从摄像机捕获截图并返回图像\n- `screenshot-game-view` - 从 Unity 编辑器的游戏视图中捕获截图\n- `screenshot-scene-view` - 从 Unity 编辑器的场景视图中捕获截图\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>脚本编写与编辑器\u003C\u002Fsummary>\n\n- `console-get-logs` - 获取 Unity 编辑器日志，并提供筛选选项\n- `editor-application-get-state` - 返回 Unity 编辑器应用程序状态信息（播放模式、暂停状态、编译状态）\n- `editor-application-set-state` - 控制 Unity 编辑器的应用程序状态（启动\u002F停止\u002F暂停播放模式）\n- `editor-selection-get` - 获取 Unity 编辑器中当前选中的信息\n- `editor-selection-set` - 设置 Unity 编辑器中的当前选中内容\n- `reflection-method-call` - 调用任意 C# 方法，传入参数并返回结果\n- `reflection-method-find` - 使用 C# 反射技术在项目中查找方法（甚至包括私有方法）\n- `script-delete` - 删除脚本文件\n- `script-execute` - 使用 Roslyn 动态编译并执行 C# 代码\n- `script-read` - 读取脚本文件的内容\n- `script-update-or-create` - 使用提供的 C# 代码更新或创建脚本文件\n- `tests-run` - 执行 Unity 测试（编辑模式\u002F播放模式），并提供筛选和详细结果\n\n\u003C\u002Fdetails>\n\n## 安装更多技能与工具\n\n当需要更多工具时，可以安装扩展；或者 [创建您自己的工具](#add-custom-tool)。\n\n| 扩展 | 描述 |\n| --- | --- |\n| **[AI 动画](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-AI-Animation\u002F)** | 一套用于 Unity 动画的附加工具 |\n| **[AI 粒子系统](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-AI-ParticleSystem\u002F)** | 一套用于 Unity 粒子系统的附加工具 |\n| **[AI ProBuilder](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-AI-ProBuilder\u002F)** | 一套用于 Unity ProBuilder 的附加工具 |\n\n![AI 游戏开发者 — Unity 技能与 MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# 目录\n\n- [快速入门](#quick-start)\n- [技能与工具参考](#skills-and-tools-reference)\n  - [安装额外的技能和工具](#install-additional-skills-and-tools)\n- [目录](#contents)\n  - [更多文档](#more-documentation)\n- [安装](#installation)\n  - [步骤1：安装`Unity MCP 插件`](#step-1-install-unity-mcp-plugin)\n    - [选项1 - 安装程序](#option-1---installer)\n    - [选项2 - CLI（推荐）](#option-2---cli-recommended)\n  - [步骤2：安装`AI代理`](#step-2-install-ai-agent)\n  - [步骤3：配置`AI代理`](#step-3-configure-ai-agent)\n    - [自动配置](#automatic-configuration)\n    - [手动配置](#manual-configuration)\n      - [命令行配置](#command-line-configuration)\n- [AI 工作流示例](#ai-workflow-examples)\n  - [LLM 的高级功能](#advanced-features-for-llm)\n    - [核心能力](#core-capabilities)\n    - [反思驱动的功能](#reflection-powered-features)\n- [自定义工具](#customize-tools)\n  - [添加自定义`Tool`](#add-custom-tool)\n  - [添加自定义`MCP Prompt`](#add-custom-mcp-prompt)\n- [运行时使用（游戏中）](#runtime-usage-in-game)\n  - [示例：AI 驱动的国际象棋游戏机器人](#sample-ai-powered-chess-game-bot)\n  - [为什么需要运行时使用？](#why-runtime-usage-is-needed)\n- [Unity `MCP 服务器` 设置](#unity-mcp-server-setup)\n  - [变量](#variables)\n  - [插件变量](#plugin-variables)\n  - [Docker 📦](#docker-)\n    - [`streamableHttp` 传输](#streamablehttp-transport)\n    - [`stdio` 传输](#stdio-transport)\n    - [自定义`端口`](#custom-port)\n  - [二进制可执行文件](#binary-executable)\n- [Unity MCP 架构的工作原理](#how-unity-mcp-architecture-works)\n  - [什么是`MCP`](#what-is-mcp)\n  - [什么是`AI代理`](#what-is-ai-agent)\n  - [什么是`MCP 服务器`](#what-is-mcp-server)\n  - [什么是`MCP Tool`](#what-is-mcp-tool)\n    - [何时使用`Tool`](#when-to-use-tool)\n  - [什么是`MCP Resource`](#what-is-mcp-resource)\n    - [何时使用`MCP Resource`](#when-to-use-mcp-resource)\n  - [什么是`MCP Prompt`](#what-is-mcp-prompt)\n    - [何时使用`MCP Prompt`](#when-to-use-mcp-prompt)\n- [贡献 💙💛](#contribution-)\n\n## 更多文档\n\n| 文档 | 描述 |\n| -------- | ----------- |\n| [默认 MCP 工具](docs\u002Fdefault-mcp-tools.md) | 包含描述的全部内置工具的完整参考 |\n| [MCP 服务器设置](docs\u002Fmcp-server.md) | 服务器配置、环境变量、远程托管 |\n| [Docker 部署](docs\u002FDOCKER_DEPLOYMENT.md) | 分步 Docker 部署指南 |\n| [开发指南](docs\u002Fdev\u002FDevelopment.md) | 架构、代码风格、CI\u002FCD — 供贡献者参考 |\n| [Wiki](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fwiki) | 入门、教程、API 参考、常见问题解答 |\n| [CLI 工具](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md) | 通过命令行安装插件、配置并连接 |\n\n![AI 游戏开发者 — Unity 技能与 MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# 安装\n\n## 步骤1：安装`Unity MCP 插件`\n\n\u003Cdetails>\n  \u003Csummary>\u003Cb>⚠️ 要求（点击）\u003C\u002Fb>\u003C\u002Fsummary>\n\n> [!IMPORTANT]\n> **项目路径不能包含空格**\n>\n> - ✅ `C:\u002FMyProjects\u002FMyProject`\n> - ❌ `C:\u002FMy Projects\u002FMyProject`\n> - ❌ `C:\u002FMy Projects\u002FMy Project`\n> - ❌ `C:\u002FMyProjects\u002FMy Project`\n\n\u003C\u002Fdetails>\n\n### 选项1 - 安装程序\n\n- **[⬇️ 下载安装程序](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Freleases\u002Flatest\u002Fdownload\u002FAI-Game-Dev-Installer.unitypackage)**\n- **📂 将安装程序导入 Unity 项目**\n  > - 您可以双击该文件——Unity 会自动打开它\n  > - 或者：先打开 Unity 编辑器，然后点击 `Assets\u002FImport Package\u002FCustom Package`，再选择该文件\n\n### 选项2 - CLI（推荐）\n\n通过 [`unity-mcp-cli`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md) 安装插件——无需 Unity 编辑器：\n\n```bash\n# 1.1 安装 unity-mcp-cli                                #  ┌────────────────────┐\nnpm install -g unity-mcp-cli                               #  │ 可用的 AI 代理 │\n                                                           #  ├────────────────────┤\n# 1.2（可选）安装 Unity                             #  │ antigravity        │\nunity-mcp-cli install-unity                                #  │ claude-code        │\n                                                           #  │ claude-desktop     │\n# 1.3（可选）创建 Unity 项目                      #  │ cline              │\nunity-mcp-cli create-project .\u002FMyUnityProject              #  │ codex              │\n                                                           #  │ cursor             │\n# 2. 在 Unity 项目中安装“AI 游戏开发者”          #  │ gemini             │\nunity-mcp-cli install-plugin .\u002FMyUnityProject              #  │ github-copilot-cli │\n                                                           #  │ kilo-code          │\n# 3. 登录云端服务器                                 #  │ open-code          │\nunity-mcp-cli login .\u002FMyUnityProject                       #  │ rider-junie        │\n                                                           #  │ unity-ai           │\n# 4. 打开 Unity 项目（自动连接并生成技能）  │ vs-copilot         │\nunity-mcp-cli open .\u002FMyUnityProject                        #  │ vscode-copilot     │\n                                                           #  └────────────────────┘\n# 5. 等待 Unity 编辑器准备就绪\nunity-mcp-cli wait-for-ready .\u002FMyUnityProject\n```\n\n> 请参阅 [完整的 CLI 文档](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fcli\u002FREADME.md)，以了解所有可用的命令。\n\n## 步骤2：安装`AI代理`\n\n请选择您偏好的单一`AI代理`——您无需安装所有代理。这将是您与 LLM 交流的主要窗口。\n\n- [Claude Code](https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fclaude-code) **（推荐）**\n- [Claude Desktop](https:\u002F\u002Fclaude.ai\u002Fdownload)\n- [GitHub Copilot in VS Code](https:\u002F\u002Fcode.visualstudio.com\u002Fdocs\u002Fcopilot\u002Foverview)\n- [Antigravity](https:\u002F\u002Fantigravity.google\u002F)\n- [Cursor](https:\u002F\u002Fwww.cursor.com\u002F)\n- [Windsurf](https:\u002F\u002Fwindsurf.com)\n- 任何其他支持的代理\n\n> AI 游戏开发者非常通用，因此您可以使用任何您喜欢的 AI 代理——它将像其他代理一样顺畅地工作。唯一重要的要求是，AI 代理必须支持技能或动态的 MCP 工具更新。\n\n## 步骤3：配置`AI代理`\n\n### 自动配置\n\n- 打开 Unity 项目\n- 打开 `Window\u002FAI 游戏开发者`\n- 选项1：点击 `自动生成` 技能 **（推荐）**\n- 选项2：点击 `配置` 模型上下文协议（MCP）\n\n![Unity_AI](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_0efd8c6c3cea.gif)\n\n> 如果您的 MCP 客户端不在列表中，请使用窗口中显示的原始 JSON 将其注入到您的 MCP 客户端中。请阅读针对您特定 MCP 客户端的说明，了解如何操作。\n\n### 手动配置\n\n如果自动配置由于任何原因无法正常工作，请使用 `AI Game Developer (Unity-MCP)` 窗口中的 JSON 文件，手动配置任何 `MCP 客户端`。\n\n#### 命令行配置\n\n\u003Cdetails>\n  \u003Csummary>\u003Cb>创建 \u003Ccode>command\u003C\u002Fcode>\u003C\u002Fb>\u003C\u002Fsummary>\n\n**1. 根据您的环境选择合适的 `\u003Ccommand>`**\n\n| 平台            | `\u003Ccommand>`                                                                                                 |\n| ------------------- | ----------------------------------------------------------------------------------------------------------- |\n| Windows x64         | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fwin-x64\u002Funity-mcp-server.exe\" port=\u003Cport> client-transport=stdio`   |\n| Windows x86         | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fwin-x86\u002Funity-mcp-server.exe\" port=\u003Cport> client-transport=stdio`   |\n| Windows arm64       | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fwin-arm64\u002Funity-mcp-server.exe\" port=\u003Cport> client-transport=stdio` |\n| MacOS Apple-Silicon | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fosx-arm64\u002Funity-mcp-server\" port=\u003Cport> client-transport=stdio`     |\n| MacOS Apple-Intel   | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fosx-x64\u002Funity-mcp-server\" port=\u003Cport> client-transport=stdio`       |\n| Linux x64           | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Flinux-x64\u002Funity-mcp-server\" port=\u003Cport> client-transport=stdio`     |\n| Linux arm64         | `\"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Flinux-arm64\u002Funity-mcp-server\" port=\u003Cport> client-transport=stdio`   |\n\n**2. 将 `\u003CunityProjectPath>` 替换为 Unity 项目的完整路径**\n\n**3. 将 `\u003Cport>` 替换为 AI Game Developer 配置中的端口号**\n\n**4. 使用命令行添加 MCP 服务器**\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_19ba6dae62cf.png\" width=\"16\" height=\"16\" alt=\"Gemini CLI\"> Gemini CLI\u003C\u002Fsummary>\n\n  ```bash\n  gemini mcp add ai-game-developer \u003Ccommand>\n  ```\n\n  > 将 `\u003Ccommand>` 替换为上表中的内容\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_c1628fa8b570.png\" width=\"16\" height=\"16\" alt=\"Claude Code CLI\"> Claude Code CLI\u003C\u002Fsummary>\n\n  ```bash\n  claude mcp add ai-game-developer \u003Ccommand>\n  ```\n\n  > 将 `\u003Ccommand>` 替换为上表中的内容\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n  \u003Csummary>\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_readme_628daf7c3c20.png\" width=\"16\" height=\"16\" alt=\"GitHub Copilot CLI\"> GitHub Copilot CLI\u003C\u002Fsummary>\n\n  ```bash\n  copilot\n  ```\n\n  ```bash\n  \u002Fmcp add\n  ```\n\n  服务器名称：`ai-game-developer`\n  服务器类型：`local`\n  命令：`\u003Ccommand>`\n  > 将 `\u003Ccommand>` 替换为上表中的内容\n\u003C\u002Fdetails>\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# AI 工作流示例\n\n与您 `AI 代理` 中的 AI（LLM）进行交流。请它完成您想要的任何任务。您对任务或想法描述得越清晰，它就能越好地完成工作。\n\n一些 `AI 代理` 允许您选择不同的 LLM 模型。请注意这一功能，因为某些模型的表现可能远优于其他模型。\n\n**示例提示：**\n\n```text\n解释我的场景层级结构\n```\n\n```text\n在顶部堆叠三个球体\n```\n\n```text\n创建金属金色材质，并将其应用到一个新的球形游戏对象上\n```\n\n> 如果使用带有 Copilot 的 VS Code，请确保已启用“代理”模式\n\n## LLM 的高级功能\n\nUnity MCP 提供了先进的工具，使 LLM 能够更快、更有效地工作，避免错误并在出现错误时自我纠正。这一切都是为了高效地实现您的目标。\n\n### 核心能力\n\n- ✔️ **适用于代理的工具** - 只需 1–2 步即可找到所需内容\n- ✔️ **即时编译** - 使用 `Roslyn` 编译和执行 C# 代码，以加快迭代速度\n- ✔️ **完全访问资源** - 对资源和 C# 脚本具有读写权限\n- ✔️ **智能反馈** - 提供清晰的正面和负面反馈，以便正确理解问题\n\n### 基于反思的功能\n\n- ✔️ **对象引用** - 提供现有对象的引用，以便快速生成 C# 代码\n- ✔️ **项目数据访问** - 以可读格式获取整个项目的数据\n- ✔️ **精细修改** - 可以填充和修改项目中的任何数据\n- ✔️ **方法发现** - 在整个代码库中查找任何方法，包括已编译的 DLL 文件\n- ✔️ **方法执行** - 调用整个代码库中的任何方法\n- ✔️ **高级参数** - 为方法调用提供任何属性，甚至内存中现有对象的引用\n- ✔️ **实时 Unity API** - Unity API 即时可用——即使 Unity 发生变化，您也能获得最新的 API\n- ✔️ **自文档化** - 通过 `Description` 属性访问任何 `类`、`方法` 或 `属性` 的人类可读描述\n\n![AI Game Developer — Unity SKILLS and MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# 自定义工具\n\n**[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)** 支持项目所有者自定义开发 `MCP 工具`、`MCP 资源` 和 `MCP 提示`。MCP 服务器会从 `Unity MCP 插件` 获取数据，并将其暴露给客户端。MCP 通信链中的任何人都将收到有关新 MCP 功能的信息，LLM 可能会在适当的时候决定使用这些功能。\n\n## 添加自定义工具\n\n要添加自定义工具，您需要：\n\n1. 一个带有 `McpPluginToolType` 属性的类\n2. 类中一个带有 `McpPluginTool` 属性的方法\n3. *可选：* 为每个方法参数添加 `Description` 属性，以帮助 LLM 理解其含义\n4. *可选：* 使用带有 `?` 和默认值的 `string? optional = null` 属性，将其标记为 LLM 的可选项\n\n> 请注意，`MainThread.Instance.Run(() =>` 这一行允许您在主线程上运行代码，这是与 Unity API 交互所必需的。如果您不需要这样做，并且可以在后台线程上运行工具，则为了避免效率损失，建议不要使用主线程。\n\n```csharp\n[McpPluginToolType]\npublic class Tool_GameObject\n{\n    [McpPluginTool\n    (\n        \"MyCustomTask\",\n        Title = \"创建一个新的游戏对象\"\n    )]\n    [Description(\"在此向 LLM 解释这是什么，以及何时应该调用它。\")]\n    public string CustomTask\n    (\n        [Description(\"向 LLM 解释这是什么。\")]\n        string inputData\n    )\n    {\n        \u002F\u002F 在后台线程中执行任何操作\n\n        return MainThread.Instance.Run(() =>\n        {\n            \u002F\u002F 如果需要，在主线程中执行一些操作\n\n            return $\"[成功] 操作已完成。\";\n        });\n    }\n}\n```\n\n## 添加自定义 `MCP 提示`\n\n`MCP 提示` 允许您将自定义提示注入到与 LLM 的对话中。它支持两种发送者角色：用户和助手。这是一种快速指导 LLM 执行特定任务的方式。您可以使用自定义数据生成提示，提供列表或其他相关信息。\n\n```csharp\n[McpPluginPromptType]\npublic static class Prompt_ScriptingCode\n{\n    [McpPluginPrompt(Name = \"add-event-system\", Role = Role.User)]\n    [Description(\"在游戏对象之间实现基于 UnityEvent 的通信系统。\")]\n    public string AddEventSystem()\n    {\n        return \"使用 UnityEvents、UnityActions 或自定义事件委托创建事件系统，以实现游戏系统和组件之间的解耦通信。\";\n    }\n}\n```\n\n![AI 游戏开发者 — Unity 技能与 MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# 运行时使用（游戏中）\n\n在您的游戏或应用中使用 **[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)**。可以使用工具、资源或提示。默认情况下没有工具，您需要实现自己的自定义工具。\n\n```csharp\n\u002F\u002F 构建 MCP 插件\nvar mcpPlugin = UnityMcpPluginRuntime.Initialize(builder =>\n    {\n        builder.WithConfig(config =>\n        {\n            config.Host = \"http:\u002F\u002Flocalhost:8080\";\n            config.Token = \"your-token\";\n        });\n        \u002F\u002F 自动注册当前程序集中的所有工具\n        builder.WithToolsFromAssembly(Assembly.GetExecutingAssembly());\n    })\n    .Build();\n\nawait mcpPlugin.Connect(); \u002F\u002F 启动与 Unity-MCP-Server 的主动连接，并自动重试\nawait mcpPlugin.Disconnect(); \u002F\u002F 停止主动连接并关闭现有连接\n```\n\n## 示例：AI 驱动的国际象棋游戏机器人\n\n这里有一个经典的国际象棋游戏。让我们将机器人逻辑外包给 LLM，由机器人根据游戏规则完成回合操作。\n\n```csharp\n[McpPluginToolType]\npublic static class ChessGameAI\n{\n    [McpPluginTool(\"chess-do-turn\", Title = \"执行回合\")]\n    [Description(\"在国际象棋游戏中执行一个回合。如果回合被接受则返回 true，否则返回 false。\")]\n    public static Task\u003Cbool> DoTurn(int figureId, Vector2Int position)\n    {\n        return MainThread.Instance.RunAsync(() => ChessGameController.Instance.DoTurn(figureId, position));\n    }\n\n    [McpPluginTool(\"chess-get-board\", Title = \"获取棋盘\")]\n    [Description(\"获取当前国际象棋棋盘的状态。\")]\n    public static Task\u003CBoardData> GetBoard()\n    {\n        return MainThread.Instance.RunAsync(() => ChessGameController.Instance.GetBoardData());\n    }\n}\n```\n\n## 为什么需要运行时使用？\n\n应用场景有很多，比如您正在开发一款带有机器人的国际象棋游戏。只需编写几行代码，就可以将机器人的决策逻辑外包给 LLM。\n\n![AI 游戏开发者 — Unity 技能与 MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# Unity `MCP 服务器` 设置\n\n**[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)** 服务器支持多种启动方式和 Docker 部署。它同时支持两种传输协议：`streamableHttp` 和 `stdio`。如果您需要自定义或在云端部署 Unity MCP 服务器，本节内容将对您有所帮助。[阅读更多...](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fmcp-server.md)\n\n## 变量\n\n无论您选择哪种启动方式，它们都支持通过环境变量和命令行参数进行自定义配置。如果您只是想简单启动服务器，使用默认值即可，无需花费时间设置这些变量。请确保 Unity 插件也使用默认值，尤其是 `--port` 参数，两者应保持一致。\n\n| 环境变量         | 命令行参数    | 描述                                                                  |\n| ---------------------------- | -------------------- | ---------------------------------------------------------------------------- |\n| `MCP_PLUGIN_PORT`            | `--port`             | **客户端** -> **服务器** \u003C- **插件** 的连接端口（默认：8080）       |\n| `MCP_PLUGIN_CLIENT_TIMEOUT`   | `--plugin-timeout`   | **插件** -> **服务器** 的连接超时时间（毫秒）（默认：10000）            |\n| `MCP_PLUGIN_CLIENT_TRANSPORT` | `--client-transport` | **客户端** -> **服务器** 的传输类型：`stdio` 或 `streamableHttp`（默认：`streamableHttp`） |\n\n> 命令行参数既支持单个 `-` 前缀的形式（如 `-port`），也支持不带前缀的形式（如 `port`）。\n> **传输方式的选择：** 当 MCP 客户端直接启动服务器二进制文件时（本地使用——这是最常见的设置），请使用 `stdio`。当服务器作为独立进程运行，或在 Docker\u002F云环境中运行并通过 HTTP 连接时，请使用 `streamableHttp`。\n\n## 插件变量\n\nUnity MCP 插件在启动时会读取以下环境变量（以及命令行参数），以覆盖保存的配置文件中的值。这些覆盖会在运行时生效；首次运行或生成新的身份验证令牌时，覆盖后的值会被**写入配置文件**。在后续运行中，覆盖值仅在内存中生效，不会自动保存。例外情况是 `UNITY_MCP_TOOLS`，它使用 `[JsonIgnore]` 注解，**永远不会持久化**——仅在运行时有效。\n\n| 环境变量        | 命令行参数            | 值              | 描述                                   |\n| --------------------------- | --------------------------- | ------------------- | --------------------------------------------- |\n| `UNITY_MCP_HOST`            | `-UNITY_MCP_HOST`           | URL 字符串                    | 覆盖 MCP 服务器的主机 URL                                    |\n| `UNITY_MCP_KEEP_CONNECTED`  | `-UNITY_MCP_KEEP_CONNECTED` | `true` \u002F `false`              | 强制启用或禁用主动连接                       |\n| `UNITY_MCP_AUTH_OPTION`     | `-UNITY_MCP_AUTH_OPTION`    | `none` \u002F `required`           | 强制设置身份验证模式                                   |\n| `UNITY_MCP_TOKEN`           | `-UNITY_MCP_TOKEN`          | 字符串                        | 强制设置身份验证令牌                                  |\n| `UNITY_MCP_TOOLS`           | `-UNITY_MCP_TOOLS`          | 逗号分隔的工具 ID      | 仅启用列出的工具；其他工具将被禁用。未知的工具 ID 将被记录为错误。 |\n\n> 命令行参数优先于环境变量。两者都会覆盖保存的配置文件中的值。\n\n**示例（CI\u002FCD 批处理模式）：**\n\n```bash\nUnity.exe -batchmode -nographics \\\n  -UNITY_MCP_HOST=http:\u002F\u002Flocalhost:8080 \\\n  -UNITY_MCP_KEEP_CONNECTED=true \\\n  -UNITY_MCP_AUTH_OPTION=required \\\n  -UNITY_MCP_TOKEN=my-secret-token\n```\n\n## Docker 📦\n\n[![Docker 镜像](https:\u002F\u002Fimg.shields.io\u002Fdocker\u002Fimage-size\u002Fivanmurzakdev\u002Funity-mcp-server\u002Flatest?label=Docker%20Image&logo=docker&labelColor=333A41 'Docker 镜像')](https:\u002F\u002Fhub.docker.com\u002Fr\u002Fivanmurzakdev\u002Funity-mcp-server)\n\n请确保已安装 Docker。如果您使用的是 Windows 操作系统，请务必启动 Docker Desktop。\n\n[阅读高级 Docker 配置说明](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002FDOCKER_DEPLOYMENT.md)。\n\n### `streamableHttp` 传输方式\n\n```bash\ndocker run -p 8080:8080 ivanmurzakdev\u002Funity-mcp-server\n```\n\n\u003Cdetails>\n  \u003Csummary>\u003Ccode>MCP 客户端\u003C\u002Fcode>配置：\u003C\u002Fsummary>\n\n```json\n{\n  \"mcpServers\": {\n    \"ai-game-developer\": {\n      \"url\": \"http:\u002F\u002Flocalhost:8080\"\n    }\n  }\n}\n```\n\n> 如果您的服务托管在云端，请将 `url` 替换为实际的端点。\n\u003C\u002Fdetails>\n\n### `stdio` 传输方式\n\n要使用此变体，`MCP 客户端` 应在 Docker 中启动 `MCP 服务器`。这可以通过修改后的 `MCP 客户端` 配置来实现。\n\n```bash\ndocker run -t -e MCP_PLUGIN_CLIENT_TRANSPORT=stdio -p 8080:8080 ivanmurzakdev\u002Funity-mcp-server\n```\n\n\u003Cdetails>\n  \u003Csummary>\u003Ccode>MCP 客户端\u003C\u002Fcode>配置：\u003C\u002Fsummary>\n\n```json\n{\n  \"mcpServers\": {\n    \"ai-game-developer\": {\n      \"command\": \"docker\",\n      \"args\": [\n        \"run\",\n        \"-t\",\n        \"-e\",\n        \"MCP_PLUGIN_CLIENT_TRANSPORT=stdio\",\n        \"-p\",\n        \"8080:8080\",\n        \"ivanmurzakdev\u002Funity-mcp-server\"\n      ]\n    }\n  }\n}\n```\n\n\u003C\u002Fdetails>\n\n### 自定义 `port`\n\n```bash\ndocker run -e MCP_PLUGIN_PORT=123 -p 123:123 ivanmurzakdev\u002Funity-mcp-server\n```\n\n\u003Cdetails>\n  \u003Csummary>\u003Ccode>MCP 客户端\u003C\u002Fcode>配置：\u003C\u002Fsummary>\n\n```json\n{\n  \"mcpServers\": {\n    \"ai-game-developer\": {\n      \"url\": \"http:\u002F\u002Flocalhost:123\"\n    }\n  }\n}\n```\n\n> 如果您的服务托管在云端，请将 `url` 替换为实际的端点。\n\u003C\u002Fdetails>\n\n## 二进制可执行文件\n\n您可以直接从二进制文件运行 Unity `MCP 服务器`。您需要拥有专门为您的 CPU 架构编译的二进制文件。请查看 [GitHub 发布页面](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Freleases)，其中包含了适用于所有 CPU 架构的预编译二进制文件。\n\n```bash\n.\u002Funity-mcp-server --port 8080 --plugin-timeout 10000 --client-transport stdio\n```\n\n\u003Cdetails>\n  \u003Csummary>\u003Ccode>MCP 客户端\u003C\u002Fcode>配置：\u003C\u002Fsummary>\n\n> 将 `\u003Cproject>` 替换为您 Unity 项目的路径。\n\n```json\n{\n  \"mcpServers\": {\n    \"ai-game-developer\": {\n      \"command\": \"\u003Cproject>\u002FLibrary\u002Fmcp-server\u002Fwin-x64\u002Funity-mcp-server.exe\",\n      \"args\": [\n        \"--port=8080\",\n        \"--plugin-timeout=10000\",\n        \"--client-transport=stdio\"\n      ]\n    }\n  }\n}\n```\n\n\u003C\u002Fdetails>\n\n![AI 游戏开发者 — Unity 技能与 MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# Unity MCP 架构的工作原理\n\n**[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)** 充当 LLM 与 Unity 之间的桥梁。它向 LLM 公开并解释 Unity 的工具，使 LLM 能够理解这些接口，并根据用户请求使用这些工具。\n\n通过集成的 `AI 连接器` 窗口，将 **[Unity MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP)** 连接到诸如 [Claude](https:\u002F\u002Fclaude.ai\u002Fdownload) 或 [Cursor](https:\u002F\u002Fwww.cursor.com\u002F) 等 LLM 客户端。同时也支持自定义客户端。\n\n该系统具有高度可扩展性——您可以在 Unity 项目代码库中直接定义自定义的 `MCP 工具`、`MCP 资源` 或 `MCP 提示词`，从而向 AI 或自动化客户端暴露新的功能。这使得 Unity MCP 成为构建高级工作流、快速原型设计以及将 AI 驱动的功能集成到开发流程中的灵活基础。\n\n## 什么是 `MCP`\n\nMCP——模型上下文协议。简而言之，它是 AI 的“USB Type-C”，特别是针对 LLM（大型语言模型）。它教会 LLM 如何使用外部功能，例如本例中的 Unity 引擎，甚至是您代码中的自定义 C# 方法。[官方文档](https:\u002F\u002Fmodelcontextprotocol.io\u002F)。\n\n## 什么是 `AI 代理`\n\n它是一个带有聊天窗口的应用程序。它可以配备智能代理以更好地运行，也可以嵌入先进的 MCP 工具。总的来说，优秀的 MCP 客户端是任务成功执行的关键因素之一，因此选择最适合的客户端非常重要。\n\n## 什么是 `MCP 服务器`\n\n它是 `MCP 客户端` 与“其他事物”之间的桥梁，在本例中就是 Unity 引擎。本项目包含 `MCP 服务器`。\n\n## 什么是 `MCP 工具`\n\n`MCP 工具` 是 LLM 可调用以与 Unity 交互的函数或方法。这些工具充当自然语言请求与实际 Unity 操作之间的桥梁。当您要求 AI“创建一个立方体”或“更改材质颜色”时，它会使用 MCP 工具来执行这些操作。\n\n**关键特性：**\n\n- **可执行的函数**，用于执行特定操作\n- **带描述的类型化参数**，帮助 LLM 理解应提供哪些数据\n- **返回值**，用于反馈操作的成功或失败\n- **线程感知**——可以在线程主线上运行 Unity API 调用，或在后台线程上进行密集处理\n\n### 何时使用 `Tool`\n\n- **自动化重复性任务**——为经常执行的常见操作创建工具\n- **复杂操作**——将多个 Unity API 调用打包成一个易于使用的工具\n- **项目特定的工作流程**——构建能够理解您项目特定结构和约定的工具\n- **易出错的任务**——创建包含验证和错误处理的工具\n- **自定义游戏逻辑**——向 AI 公开您的游戏系统，以实现动态内容生成\n\n**示例：**\n\n- 创建并配置带有特定组件的 GameObject\n- 批量处理资源（纹理、材质、预制件）\n- 设置光照和后期处理效果\n- 生成关卡几何体或程序化放置物体\n- 配置物理设置或碰撞层\n\n## 什么是 `MCP 资源`\n\n`MCP 资源` 提供对 Unity 项目内数据的只读访问。与执行操作的 MCP 工具不同，资源允许 LLM 检查和理解您项目的当前状态、资产和配置。可以将其视为向 AI 提供项目上下文的“传感器”。\n\n**关键特性：**\n\n- **只读访问**项目数据和 Unity 对象\n- **结构化信息**，以 LLM 可理解的格式呈现\n- **实时数据**，反映您项目的当前状态\n- **情境意识**，帮助 AI 做出明智决策\n\n### 何时使用 `MCP Resource`\n\n- **项目分析**：让 AI 理解您的项目结构、资源和组织方式。\n- **调试辅助**：提供当前状态信息以帮助排查问题。\n- **智能建议**：为 AI 提供上下文，以便做出更好的推荐。\n- **文档生成**：根据项目状态自动生成文档。\n- **资产管理**：帮助 AI 了解可用的资源及其属性。\n\n**示例：**\n\n- 暴露场景层级和 GameObject 属性。\n- 列出可用的材质、纹理及其设置。\n- 显示脚本依赖关系和组件关联。\n- 展示当前的光照设置和渲染管线配置。\n- 提供关于音频源、动画和粒子系统的信息。\n\n## 什么是 `MCP Prompt`\n\n`MCP Prompt` 允许您将预定义的提示注入到与 LLM 的对话中。这些是智能模板，可以提供上下文、指令或知识，以引导 AI 的行为。提示可以是静态文本，也可以根据项目的当前状态动态生成。\n\n**关键特性：**\n\n- **上下文引导**，影响 AI 的响应方式。\n- **角色化**——可模拟不同角色（用户请求或助手知识）。\n- **动态内容**——可包含实时项目数据。\n- **可重用模板**——适用于常见场景和工作流。\n\n### 何时使用 `MCP Prompt`\n\n- **提供领域知识**：分享针对您项目的最佳实践和编码规范。\n- **设定编码规范**：确立命名规范、架构模式和代码风格。\n- **提供项目结构背景**：解释您的项目是如何组织的以及原因。\n- **分享工作流程说明**：为常见任务提供分步操作指南。\n- **注入专业知识**：添加关于特定 Unity 功能、第三方资源或自定义系统的信息。\n\n**示例：**\n\n- “公共方法始终使用 PascalCase，私有字段则使用 camelCase。”\n- “该项目使用位于 Scripts\u002FEvents\u002F 的自定义事件系统。”\n- “创建 UI 元素时，务必将其添加到 Scene\u002FUI\u002FMainCanvas 中的 Canvas 上。”\n- “性能至关重要——对于频繁实例化的对象，优先使用对象池技术。”\n- “该项目遵循 SOLID 原则——请解释任何架构决策。”\n\n![AI 游戏开发者 — Unity 技能与 MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)\n\n# 贡献 💙💛\n\n我们非常欢迎您的贡献！请带来您的想法，让我们一起把游戏开发变得比以往任何时候都更加简单吧！您是否有关于新工具或功能的想法，或者发现了一个 bug 并知道如何修复它？\n\n**如果您觉得这个项目有用，请给它点个赞 🌟！**\n\n1. 👉 [阅读开发文档](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fdev\u002FDevelopment.md)\n2. 👉 [Fork 该项目](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Ffork)\n3. 克隆您的分支，并在 Unity 中打开 `.\u002FUnity-MCP-Plugin` 文件夹。\n4. 在项目中实现新功能，提交更改并推送到 GitHub。\n5. 创建一个 Pull Request，目标仓库为原始的 [Unity-MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare) 仓库，分支为 `main`。\n\n![AI 游戏开发者 — Unity 技能与 MCP](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fblob\u002Fmain\u002Fdocs\u002Fimg\u002Fpromo\u002Fhazzard-divider.svg?raw=true)","# Unity-MCP 快速上手指南\n\nUnity-MCP 是一款基于模型上下文协议（MCP）的 AI 游戏开发助手，支持在 **Unity 编辑器**和**编译后的游戏运行时**中使用。它允许你将 Claude、Cursor、GitHub Copilot 等 AI 代理连接到 Unity，实现自动化工作流、代码生成以及游戏内的动态 AI 交互。\n\n## 环境准备\n\n在开始之前，请确保满足以下要求：\n\n*   **操作系统**：Windows, macOS 或 Linux。\n*   **Unity 编辑器**：支持主流 Unity 版本（具体兼容版本请参考官方发布页）。\n*   **Node.js**：需安装 Node.js 以使用 CLI 工具（推荐 LTS 版本）。\n*   **AI 客户端**：已安装并配置好任意支持 MCP 的 AI 工具（如 Claude Desktop, Cursor, Windsurf, GitHub Copilot 等）。\n*   **网络环境**：由于依赖 npm 和 OpenUPM，国内用户建议配置淘宝镜像或科学网络环境以确保下载速度。\n\n## 安装步骤\n\n你可以通过 **CLI 命令行** 或 **手动安装包** 两种方式将 Unity-MCP 集成到项目中。\n\n### 方法一：使用 CLI 安装（推荐）\n\n这是最快捷的方式，可自动处理依赖和配置。\n\n1.  **安装 CLI 工具**\n    在终端中运行以下命令全局安装 `unity-mcp-cli`：\n    ```bash\n    npm install -g unity-mcp-cli\n    ```\n    > *国内加速提示：若下载缓慢，可先配置 npm 镜像源：`npm config set registry https:\u002F\u002Fregistry.npmmirror.com`*\n\n2.  **在 Unity 项目中安装插件**\n    进入你的 Unity 项目根目录，执行：\n    ```bash\n    unity-mcp-cli install-plugin .\u002FMyUnityProject\n    ```\n    *(请将 `.\u002FMyUnityProject` 替换为你实际的项目路径)*\n\n3.  **登录云服务器（可选但推荐）**\n    用于生成针对当前项目的专属 AI 技能（Skills）：\n    ```bash\n    unity-mcp-cli login .\u002FMyUnityProject\n    ```\n\n4.  **打开项目并自动生成技能**\n    此命令会自动启动 Unity 编辑器并连接 MCP 服务：\n    ```bash\n    unity-mcp-cli open .\u002FMyUnityProject\n    ```\n\n### 方法二：手动安装 (.unitypackage)\n\n如果不方便使用 CLI，可直接下载安装包：\n\n1.  前往 [Releases 页面](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Freleases) 下载最新的 `AI-Game-Dev-Installer.unitypackage`。\n2.  在 Unity 编辑器中，点击菜单栏 `Assets` > `Import Package` > `Custom Package`。\n3.  选择下载的文件并导入所有资源。\n4.  导入完成后，重启 Unity 编辑器。\n\n## 基本使用\n\n安装完成后，只需三个简单步骤即可让 AI 控制你的 Unity 项目。\n\n### 第一步：配置 AI 代理\n\n1.  打开 Unity 编辑器，在顶部菜单栏找到并点击 **`Window` > `AI Game Developer`**。\n2.  在打开的面板中，点击 **`Auto-generate skills`**（自动生成技能）。\n    *   *系统会根据你的操作系统、Unity 版本及已安装的插件，为 AI 生成专属的操作能力。*\n3.  或者点击 **`Configure MCP`** 进行手动配置（适用于高级用户）。\n\n### 第二步：连接 AI 客户端\n\n确保你的 AI 编辑器（如 Cursor 或 Claude Desktop）已配置好 MCP Server 连接。\n*   如果使用 CLI 安装的 `open` 命令，服务通常会自动在本地启动（stdio 模式）。\n*   如果在远程运行，请确保 HTTP 地址配置正确。\n\n### 第三步：开始对话\n\n现在，你可以直接在 AI 聊天窗口中用自然语言下达指令。\n\n**示例指令：**\n> \"Create 3 cubes in a circle with radius 2\"\n> （在半径为 2 的圆上创建 3 个立方体）\n\n**预期结果：**\nAI 将调用内置的 `gameobject-create` 和 `gameobject-modify` 等工具，自动在场景中生成并排列好物体。\n\n---\n\n**核心功能提示：**\n*   **资源管理**：可指令 AI 创建文件夹、移动资产、修改材质球（如：`assets-create-folder`, `assets-material-create`）。\n*   **场景操作**：可指令 AI 添加组件、修改对象属性、截图（如：`gameobject-component-add`, `screenshot-game-view`）。\n*   **运行时调试**：该插件独特支持在**编译后的游戏包体**中运行，可用于实时调试 NPC 行为或让玩家与 AI 互动。","独立游戏开发者阿明正在为一款 3D 平台跳跃游戏快速迭代核心机制，需要频繁调整角色移动逻辑并验证物理效果。\n\n### 没有 Unity-MCP 时\n- **上下文割裂严重**：AI 助手无法直接读取 Unity 项目中的场景层级、Prefab 配置或 Inspector 属性，只能基于模糊描述生成通用代码，导致大量“幻觉”错误。\n- **手动测试循环冗长**：每次修改代码后，必须切回 Unity 编辑器手动编译、运行游戏、操控角色测试，再返回 IDE 记录问题，单次验证耗时数分钟。\n- **工具集成门槛高**：若想将自定义的 C# 调试方法（如“打印当前帧率”或“重置玩家位置”）暴露给 AI，需编写复杂的桥接代码或插件，开发成本高。\n- **Token 消耗巨大**：为了让 AI 理解项目结构，不得不反复粘贴大量脚本文件和配置文件内容，迅速耗尽对话窗口的上下文限额。\n\n### 使用 Unity-MCP 后\n- **实时感知项目状态**：Unity-MCP 让 Claude Code 或 Cursor 直接“看见”当前场景中的 GameObject 和组件数据，生成的代码能精准匹配现有的变量名与架构。\n- **一键执行与验证**：通过 CLI 指令，AI 可直接调用 Unity 内部方法（如 `Player.ResetPosition()`），在编辑器中即时运行测试并返回结果，将验证周期缩短至秒级。\n- **零代码暴露工具**：只需在任意 C# 方法上添加一行特性标记，即可将其转化为 AI 可调用的工具，瞬间扩展了 AI 对引擎的控制能力。\n- **高效令牌利用**：AI 按需查询引擎状态而非读取全量文件，大幅减少 Token 占用，使长对话中能处理更复杂的逻辑重构任务。\n\nUnity-MCP 彻底打通了 AI 模型与 Unity 引擎间的壁垒，将原本割裂的“编码 - 测试”流程转变为流畅的自然语言交互闭环。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FIvanMurzak_Unity-MCP_8cea62cb.png","IvanMurzak","Ivan Murzak","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FIvanMurzak_bbe5a567.png","AI inference. Making AI better. Integrating AI into production pipelines. Developing AI agents. Working with 3D Graphics and real-time physics. Full stack, 0-1.","Mirrorscape","Redmond, WA, United States","Ivan.D.Murzak@gmail.com","https:\u002F\u002Fivan-murzak.com","https:\u002F\u002Fgithub.com\u002FIvanMurzak",[86,90,94,98,102,106,110,113],{"name":87,"color":88,"percentage":89},"C#","#178600",86.5,{"name":91,"color":92,"percentage":93},"TypeScript","#3178c6",9,{"name":95,"color":96,"percentage":97},"PowerShell","#012456",3.3,{"name":99,"color":100,"percentage":101},"Python","#3572A5",1,{"name":103,"color":104,"percentage":105},"Shell","#89e051",0.2,{"name":107,"color":108,"percentage":109},"Dockerfile","#384d54",0,{"name":111,"color":112,"percentage":109},"Batchfile","#C1F12E",{"name":114,"color":115,"percentage":109},"JavaScript","#f1e05a",1898,179,"2026-04-03T10:33:06","Apache-2.0","Windows, macOS, Linux","未说明","未说明 (取决于 Unity 编辑器及运行项目的系统要求)",{"notes":124,"python":125,"dependencies":126},"该工具是 Unity 的插件和 MCP 服务器，主要通过 npm 包 (unity-mcp-cli) 或 Docker 容器进行部署。它需要安装 Unity 编辑器才能发挥完整功能，同时也支持在编译后的游戏运行时中使用。配置 AI 代理（如 Claude, Cursor 等）需要用户自行拥有相应的 API 密钥或服务订阅。未提及具体的 GPU 显存或 CUDA 版本要求，因为计算负载主要取决于所连接的外部大模型服务或本地运行的 LLM 配置，而非插件本身。","未说明 (通过 Node.js\u002Fnpm 或 Docker 部署，非纯 Python 环境)",[127,128,129],"Unity Editor (版本未限定，支持存档中的多个版本)","Node.js \u002F npm (用于安装 unity-mcp-cli)","Docker (可选，用于运行服务器)",[15,14,13,26,53],[132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,150],"ai","ai-integration","game-development","mcp","mcp-server","unity","claude","claude-code","claude-desktop","copilot","cursor","gemini","llm","model-context-protocol","openai","windsurf","gamedev","anthropic","cli",null,"2026-03-27T02:49:30.150509","2026-04-06T11:30:52.280128",[155,160,165,170,175,180],{"id":156,"question_zh":157,"answer_zh":158,"source_url":159},10470,"如何在 Windows 上将 Unity-MCP 添加到 Claude Code？","版本 0.17.2 及以上已集成配置按钮，无需 WSL。你也可以使用命令行手动添加：\n1. 将 `\u003CunityProjectPath>` 替换为你的实际项目路径。\n2. 运行以下命令（注意路径末尾需加 `.exe`）：\n`claude mcp add Unity-MCP \"\u003CunityProjectPath>\u002FLibrary\u002Fmcp-server\u002Fwin-x64\u002Funity-mcp-server.exe\" client-transport=stdio`\n\n如果使用 WSL2 Ubuntu 环境，命令如下：\n`claude mcp add-json Unity-MCP \"{ \\\"command\\\": \\\"$(pwd)\u002FLibrary\u002Fmcp-server\u002Fwin-x64\u002Funity-mcp-server.exe\\\", \\\"args\\\": [ \\\"--port=8080\\\", \\\"--timeout=120000\\\" ] }\" --scope local`","https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fissues\u002F172",{"id":161,"question_zh":162,"answer_zh":163,"source_url":164},10471,"安装后出现 'Immutable' 命名空间不存在的编译错误怎么办？","这是一个已知问题，通常发生在特定 Unity 版本（如 Unity 6）或 macOS M1 环境下。该问题已在版本 0.41.1 中修复。请确保将 Unity-MCP 包更新到 0.41.1 或更高版本即可解决。如果问题仍然存在，请检查是否遗漏了依赖项安装。","https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fissues\u002F160",{"id":166,"question_zh":167,"answer_zh":168,"source_url":169},10472,"如何使用 MCP 工具截取相机画面并发送给 LLM？","MCP 工具支持直接返回图像的 Base64 二进制数据，无需保存到临时文件夹。你可以使用现有的 Prompt API 调用截图功能。实现逻辑是捕获主相机（Main Camera）的画面，将其转换为 Base64 格式并作为图像类型内容发送给 LLM。相关代码实现可参考 PR #26，确保遵循标准协议规则，直接返回图像源数据而非文件路径。","https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fissues\u002F253",{"id":171,"question_zh":172,"answer_zh":173,"source_url":174},10473,"如何在 Unity 编辑器中添加一个持久化的按钮来打开 AI Game Developer 窗口？","Unity 官方使用 `[MainToolbarElement]` 和内部的 `[UnityOnlyMainToolbarPreset]` 属性来默认显示工具栏按钮。由于后者是内部属性，第三方插件无法直接使用。建议的替代方案是：对于所有 Unity 版本，可以将入口放在 SceneView 中；对于 Unity 6.3+ 版本，可以尝试通过公共 API 添加到工具栏。目前最稳定的方法是通过脚本在编辑器窗口中创建自定义入口。","https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fissues\u002F610",{"id":176,"question_zh":177,"answer_zh":178,"source_url":179},10474,"序列化包含 TextMeshProUGUI 的 UI 层级时报错怎么办？","当使用 MCP 工具序列化包含 TextMeshProUGUI 组件的 GameObject 层级时，可能会遇到关于材质序列化和缺少 '_Color' 属性的警告或错误（例如 'TextMeshPro\u002FDistance Field' 着色器）。目前这些错误通常不会导致工具调用失败，Agent 仍能收到部分响应。建议忽略这些非阻断性警告，或者在后续版本中关注针对 TMP 材质序列化的优化处理。","https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fissues\u002F391",{"id":181,"question_zh":182,"answer_zh":183,"source_url":184},10475,"为什么需要保留启动 GIF 动画文档？","虽然某些高级 AI 代理（如 Codex）可以直接理解上下文而不需要视觉引导，但仍有大量 AI Agent 依赖视觉示例来理解操作流程。因此，保留展示当前 UI 窗口操作的简短 GIF 动画对于文档的通用性和易用性非常重要，能帮助用户和各类 AI 快速上手。","https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fissues\u002F540",[186,191,196,201,206,211,216,221,226,231,236,241,246,251,256,261,266,271,276,281],{"id":187,"version":188,"summary_zh":189,"released_at":190},71031,"0.63.3","# AI Game Developer (Unity MCP) 0.63.3\n**Released:** *April  1, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.63.2...0.63.3](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.63.2...0.63.3)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`e3fd3060`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fe3fd3060ee7921c1c5b47d80f821427c73edf4da) — Bump version to 0.63.3 by @IvanMurzak\n- [`846dcedf`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F846dcedf4b75b099a64b5d5fcbf063eee652aef5) — fix: run console-clear-logs tool on main thread (#637) by @IvanMurzak\n- [`444f6367`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F444f6367df224d585b5e9e66557552fe8aadbec5) — feat: update Claude Code configurator YouTube tutorial URL (#636) by @IvanMurzak\n- [`ee742e14`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fee742e14ae9493e32a2915ca05d2af04f6d6ae2d) — feat: add setup skills image to documentation by @IvanMurzak","2026-04-01T06:56:48",{"id":192,"version":193,"summary_zh":194,"released_at":195},71032,"0.63.2","# AI Game Developer (Unity MCP) 0.63.2\n**Released:** *March 31, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.63.1...0.63.2](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.63.1...0.63.2)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`46688bde`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F46688bdebdc43da05d1dc0abc0cc3cadcfb1a00a) — Bump version to 0.63.2 by @IvanMurzak\n- [`81d4401f`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F81d4401f6a9a2b849b4fb32c7930633b5711e5cc) — Update McpPlugin DLLs by @IvanMurzak\n- [`6bf454c5`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F6bf454c59c8f13f1aedc24a249e20c21012556d1) — Update McpPlugin netstandard2.1 DLLs (#634) by @IvanMurzak","2026-03-31T02:43:10",{"id":197,"version":198,"summary_zh":199,"released_at":200},71033,"0.63.1","# AI Game Developer (Unity MCP) 0.63.1\n**Released:** *March 30, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.63.0...0.63.1](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.63.0...0.63.1)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`d1fb8a81`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fd1fb8a810cf896a11dc5bc2bd6bb5635e43f9156) — Bump version to 0.63.1 by @IvanMurzak\n- [`d1a6e811`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fd1a6e8110754d50c1d866ad21b9e19cccc89956d) — fix: add missing .meta files for Overlay folder and toolbar overlay scripts (#628) by @cagilfirat","2026-03-30T00:16:27",{"id":202,"version":203,"summary_zh":204,"released_at":205},71034,"0.63.0","# AI Game Developer (Unity MCP) 0.63.0\n**Released:** *March 29, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.62.0...0.63.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.62.0...0.63.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`e821055b`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fe821055b924d955b29305bf61c9bd4995721a3da) — Bump version to 0.63.0 by @IvanMurzak\n- [`365cda8d`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F365cda8d2d9f37b5f05c8c22d351e16a04fc8989) — Feature\u002Fenv variable for host url override (#626) by @IvanMurzak\n- [`a42ed0e4`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fa42ed0e490b984ce81e57c3553a29492a15ff667) — Create SKILL.md by @IvanMurzak\n- [`effa1249`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Feffa1249473504318a55fa79b46efc988740cd3d) — feat: add SceneView toolbar overlay with open-window button and tests (#616) by @cagilfirat","2026-03-29T22:17:02",{"id":207,"version":208,"summary_zh":209,"released_at":210},71035,"0.62.0","# AI Game Developer (Unity MCP) 0.62.0\n**Released:** *March 28, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.61.0...0.62.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.61.0...0.62.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`86b27ecb`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F86b27ecb0e30001e420edfcd634b9ad6d7edd7c9) — Bump project version to 0.62.0 by @IvanMurzak\n- [`12e98286`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F12e982866e93adff83a081e85ceb24258d98df89) — Bump McpPlugin.Server to 5.10.0 and update DLLs (#621) by @IvanMurzak\n- [`7de746f7`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F7de746f70d812a4f8a9f320e34dbd7e1260497c3) — ci: add disk cleanup step to Unity Plugin test workflow (#620) by @cagilfirat\n- [`4a6c6395`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F4a6c6395922a3fadb096f88615e7c66a0f436827) — feat: add tool-list MCP tool with regex search and tests (#617) by @cagilfirat","2026-03-28T05:07:15",{"id":212,"version":213,"summary_zh":214,"released_at":215},71036,"0.61.0","# AI Game Developer (Unity MCP) 0.61.0\n**Released:** *March 25, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.60.0...0.61.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.60.0...0.61.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`a5069395`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fa5069395b6adec146952fbc69dac8d0f8c0ec474) — Bump version to 0.61.0 by @IvanMurzak\n- [`6e3cf9d3`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F6e3cf9d3fef75d76bdbc1f8e2d0de1729eccfa29) — feat(cli): add self-update notification and update command (#611) by @IvanMurzak\n- [`b1c5f031`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fb1c5f031854a1a9ce4e4d4e5dca6b54b816462e9) — fix(cli): send relative path in setup-skills instead of absolute (#604) by @IvanMurzak\n- [`c98fc448`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc98fc4489b539d4b68ea387d3967aae118adbe63) — fix: make ToolSetEnabledStateTests resilient to editor config state (#605) by @cagilfirat\n- [`aca98cc0`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Faca98cc0bebeda665329286e1bdef3c6878b7b4d) — feat: add shader get-data MCP tool (#603) by @IvanMurzak\n- [`9fff7c7e`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F9fff7c7e8429dcc03ac5c74dd1876a320a78e96a) — fix: screenshot tests fail when LogLevel is None (#606) by @IvanMurzak\n- [`b27cafd4`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fb27cafd464c18d1a336792888fe618858847e504) — feat: add Gradient serialization support and improve parameter validation (#602) by @IvanMurzak","2026-03-25T23:45:44",{"id":217,"version":218,"summary_zh":219,"released_at":220},71037,"0.60.0","# AI Game Developer (Unity MCP) 0.60.0\n**Released:** *March 24, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.59.0...0.60.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.59.0...0.60.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`c754c44c`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc754c44c7d183cc07ee416359cddf8af2228f074) — Bump version to 0.60.0 by @IvanMurzak\n- [`343e3e0c`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F343e3e0cad304485d1d9fa874ccbd1e99a3f2c35) — Add tool-set-enabled-state skill docs by @IvanMurzak\n- [`3b1e6e7f`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F3b1e6e7f9495c052c012c0836d7176b406a352eb) — feat: add tool-set-enabled-state MCP tool and reactive UI updates (#592) by @cagilfirat\n- [`0b4cd248`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F0b4cd24808bc1d0b07382b6c4bc27e9c9926af7a) — refactor: implement Tool_Assets.Refresh with ResponseCallTool and compilation handling (#597) by @IvanMurzak\n- [`f2120deb`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Ff2120deb2a69abf65cf5e3c2715b4af466b73cf3) — fix: update stdin example in ping skill documentation by @IvanMurzak","2026-03-24T23:32:32",{"id":222,"version":223,"summary_zh":224,"released_at":225},71038,"0.59.0","# AI Game Developer (Unity MCP) 0.59.0\n**Released:** *March 24, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.58.0...0.59.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.58.0...0.59.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`8316f2e0`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F8316f2e035b601d0da516d25e29437917358ea25) — Bump version to 0.59.0 by @IvanMurzak\n- [`a2bf176b`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fa2bf176bf783798823b820b9d1b8fe2d4836b854) — feat: dev-loop testing infrastructure (#596) by @IvanMurzak","2026-03-24T08:42:35",{"id":227,"version":228,"summary_zh":229,"released_at":230},71039,"0.58.0","# AI Game Developer (Unity MCP) 0.58.0\n**Released:** *March 24, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.57.3...0.58.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.57.3...0.58.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`b5e63e20`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fb5e63e20b472b2870f0190416a23cb7707d36b70) — chore: bump version to 0.58.0 across all components by @IvanMurzak\n- [`c79fe0c2`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc79fe0c28906c9b6e7d1e6b4c8b45bcef7c937e1) — feat: Extensions UI section in AI Game Developer window (#584) by @IvanMurzak\n- [`a7c7f409`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fa7c7f409607055c351d6767f685c36e2ebc35015) — Add reusable AlertPanel component and connection auth alert (#594) by @IvanMurzak","2026-03-24T04:14:56",{"id":232,"version":233,"summary_zh":234,"released_at":235},71040,"0.57.3","# AI Game Developer (Unity MCP) 0.57.3\n**Released:** *March 23, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.57.2...0.57.3](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.57.2...0.57.3)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`ddb32f17`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fddb32f1763ef3e127abd270a4431f04409ce846c) — chore: bump version to 0.57.3 across all components by @IvanMurzak\n- [`99c093a1`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F99c093a17c573b014390bfc3e8360fe7aa51ae4a) — Use https:\u002F\u002Fai-game.dev for CloudServerBaseUrl by @IvanMurzak\n- [`a0aece7f`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fa0aece7f1e4e1f4dfa440c8d7d4a4e532cfdcfbf) — fix: update image links in README files for localized versions by @IvanMurzak","2026-03-23T02:45:59",{"id":237,"version":238,"summary_zh":239,"released_at":240},71041,"0.57.2","# AI Game Developer (Unity MCP) 0.57.2\n**Released:** *March 22, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.57.1...0.57.2](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.57.1...0.57.2)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`e0353d15`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fe0353d15f44ab6992e43d95a260b51980d3e7f41) — chore: bump version to 0.57.2 across all components by @IvanMurzak\n- [`83acafdb`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F83acafdb69b376e775c8518746c86d45c08e858a) — fix: update language links in README files to include English by @IvanMurzak\n- [`69f35eb2`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F69f35eb2963f85d61c5fe538b317ae0cc65f3c63) — Remove README.md file from the documentation directory by @IvanMurzak\n- [`222306a1`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F222306a12efebe84c3a59c85118348fd56a69cd2) — fix: update video links to new asset URLs in README files by @IvanMurzak\n- [`a178a39e`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fa178a39eb8ae510d579aabbf0f93cbb647e4d13b) — Update README.zh-CN.md with new branding and improved descriptions; add new video and image assets for AI Game Developer; update existing image files. by @IvanMurzak\n- [`c7bd5e42`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc7bd5e42bb362f8bd3372e89dcf2f6cf415a6f78) — fix: stop MCP server before installing package update (#591) by @i-snyder\n- [`94ec53a5`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F94ec53a5f9a54a2d8150225ca92a73997d224df5) — feat: update McpPlugin.Server package reference to version 5.5.2 by @IvanMurzak","2026-03-22T09:54:25",{"id":242,"version":243,"summary_zh":244,"released_at":245},71042,"0.57.1","# AI Game Developer (Unity MCP) 0.57.1\n**Released:** *March 20, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.57.0...0.57.1](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.57.0...0.57.1)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`86c4b9b9`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F86c4b9b976228dd02390b6052551faeb1989d505) — feat: update version to 0.57.1 across all relevant files by @IvanMurzak\n- [`501a3671`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F501a367121543962db0adfa9cf3a069b03b20e89) — feat: add setup alert panel for AI agent configuration (#590) by @IvanMurzak","2026-03-20T23:44:05",{"id":247,"version":248,"summary_zh":249,"released_at":250},71043,"0.57.0","# AI Game Developer (Unity MCP) 0.57.0\n**Released:** *March 20, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.56.0...0.57.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.56.0...0.57.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`c0f9348c`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc0f9348c874c3fb6d75a72f5deb36287e7ee91cb) — feat: add tooltips for connection status indicators in MainWindowEditor by @IvanMurzak\n- [`e7db0a4b`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fe7db0a4b98186dfec88e147587bf744c24de86b6) — feat: update version to 0.57.0 across all relevant files by @IvanMurzak\n- [`aeed96cb`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Faeed96cb2efc73b23b215dce63604a0a91e5612d) — feat: enable McpPlugin.Server package reference in project file by @IvanMurzak\n- [`d2c51e1e`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fd2c51e1ecaffda47b75adcfc1ac4ca4a43ef3b8f) — chore: remove Skills.meta file by @IvanMurzak\n- [`a346988c`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fa346988c101ee4e59dda0cdb610935e644a3b957) — Minor (#588) by @IvanMurzak\n- [`09a40567`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F09a405676c81adffb79826c8045a89a3d5c1d94f) — feat: add CLI-driven AI agent setup, skill tools, and initial setup guide (#587) by @IvanMurzak\n- [`d52eacbb`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fd52eacbb1abe4a031ad4bd18ea2270199ff0a0a5) — feat: add setup-mcp and setup-skills CLI commands (#585) by @IvanMurzak\n- [`75bcd6ac`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F75bcd6acfab899d1843deda721ea700828cebe87) — ci: set Claude Code model to claude-opus-4-6 by @IvanMurzak","2026-03-20T19:50:48",{"id":252,"version":253,"summary_zh":254,"released_at":255},71044,"0.56.0","# AI Game Developer (Unity MCP) 0.56.0\n**Released:** *March 18, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.55.1...0.56.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.55.1...0.56.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`c502cfc6`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc502cfc6623dd7cbd16e4836b8bd25ed10fa7b39) — chore: update version to 0.56.0 and enhance description in package.json by @IvanMurzak\n- [`f9d438b7`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Ff9d438b777087aeb29a320ca244ae414d3b64354) — UI: Segmented controls, styled forms, and agent config redesign (#582) by @IvanMurzak\n- [`54d7c1b0`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F54d7c1b0b43198bc1e98f98deb8c561dbf250378) — Update unity-editor.ts (#579) by @IvanMurzak\n- [`c8d13666`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc8d136664a6817eb74f8e1235736ee33a388ffaf) — Support .app paths and parse Hub output (#576) by @IvanMurzak\n- [`9dbf348a`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F9dbf348afc364b3d68fd4bdaa48dd3c9d7e31699) — fix: fixed link to kilocode (#577) by @mrwellmann\n- [`fa5072a3`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Ffa5072a30c433aa7320cce99c9b03a2eb19fa172) — Add new skills for scene management, screenshot capture, script handling, and testing by @IvanMurzak\n- [`064459d5`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F064459d5375270912bd7774198d0793296dd2c7d) — fix: change port range from 50000-59999 to 20000-29999 to avoid Windows reserved ports (#575) by @IvanMurzak","2026-03-18T20:33:14",{"id":257,"version":258,"summary_zh":259,"released_at":260},71045,"0.55.1","# AI Game Developer (Unity MCP) 0.55.1\n**Released:** *March 17, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.55.0...0.55.1](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.55.0...0.55.1)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`7a1baf1f`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F7a1baf1fe475ad82aa52dcbc07c02dab0e1b849d) — chore: bump version to 0.55.1 across all relevant files by @IvanMurzak\n- [`e1b42663`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fe1b42663237068f8002a2e489e690751d22ba76f) — fix: update command to use npx for running unity-mcp-cli in BuildToolCommand by @IvanMurzak\n- [`a4bbfffa`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fa4bbfffa08fd9f919f4e68237e31cf26308c856c) — docs: fix Docsify setup - add local README copy with relative URLs by @IvanMurzak\n- [`ece1eae2`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fece1eae2982f9a053fac85ddcec02c00ec543e0b) — docs: add Docsify-based GitHub Pages site by @IvanMurzak","2026-03-17T04:34:21",{"id":262,"version":263,"summary_zh":264,"released_at":265},71046,"0.55.0","# AI Game Developer (Unity MCP) 0.55.0\n**Released:** *March 16, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.54.0...0.55.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.54.0...0.55.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`1ffa35b8`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F1ffa35b8805c9676ccc8afdefb5e8fa9a3d864c8) — Bump version to 0.55.0 by @IvanMurzak\n- [`8be8a5d6`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F8be8a5d679913e086e5d9d289621293e074ffd8e) — refactor: update skill agent ID retrieval and change access modifier for selectedAiAgentId by @IvanMurzak\n- [`e55c347d`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fe55c347da7008a861306421a105f8f5b71840d74) — refactor: change access modifier of selectedAiAgentId to internal in MainWindowEditor by @IvanMurzak\n- [`3624268f`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F3624268f9683d9c7f8e610328ce995b596d4afe6) — feat: add run-tool CLI command (#573) by @IvanMurzak\n- [`ffa83f19`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fffa83f195a87b553d7fb7969508a464760656422) — fix: update label text for auto-generate toggle in TemplateSkillsSection by @IvanMurzak\n- [`021d7bc9`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F021d7bc9af53d81a1c3d1d2bdb947372693e803b) — feat: implement per-agent skill file auto-generation and configuration by @IvanMurzak\n- [`713c2c52`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F713c2c52dd655ef8f8024411c72036dc0d1255a3) — feat: add SkillsPath property for various agent configurators and update UXML label text by @IvanMurzak\n- [`f953c1c5`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Ff953c1c5dfa1ce035dc15ed722f33da9910bdbfc) — fix: adjust margin styles for TemplateSkillsSection in UXML by @IvanMurzak\n- [`f545a6ab`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Ff545a6abf4a2bd05370bd56c8689ba22609b5105) — feat: implement skills UI section with auto-generate toggle and generate button by @IvanMurzak\n- [`23f288a3`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F23f288a3422a3d2dc7c451e3cebb27573f85791d) — Add per-agent Skills UI and SkillsPath support by @IvanMurzak\n- [`4c7815cd`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F4c7815cd1bc10b30feac03aeb35a5dca6bd78aae) — Update README.md by @IvanMurzak\n- [`88e2cc3e`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F88e2cc3e6671bf400948877c684a08c20d327a34) — fix: correct exposure of --token option and update authorization header messages in run-tool tests by @IvanMurzak\n- [`192c05f9`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F192c05f934fe2ec54f1d70385fc4ef9303d85c7c) — fix: update asset copy descriptions for clarity and improve UnitySkillFileGenerator CLI command formatting by @IvanMurzak\n- [`1e7035cb`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F1e7035cb00499d599e0ef07a896b38302cd2a9e6) — fix: always include auth token in run-tool regardless of --url flag by @IvanMurzak\n- [`a13a665a`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fa13a665a3fe5d2531b22ec1b88e8caea2913844f) — fix: add missing using directive for McpPlugin in UnitySkillFileGenerator by @IvanMurzak\n- [`8d8c16b6`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F8d8c16b66931c22fdd98e7bd613a1dbf438f3000) — fix: address PR review comments on run-tool command by @IvanMurzak\n- [`cd264685`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fcd26468595b60379f24dccb540f8f94e29c26ae8) — feat: set default AI agent to Claude Code during initial setup by @IvanMurzak\n- [`35ba6a19`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F35ba6a199f3065c3f7d471ae2a37bda45b75a3b1) — feat: set default skills path for Claude Code in UnityMcpPlugin.Config by @IvanMurzak\n- [`5b360120`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F5b36012055e9c0a14e6e317120ecfcbc7cbaeed4) — fix: update McpPlugin.Server package version to 5.3.2 by @IvanMurzak\n- [`5efb365f`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F5efb365f614bc21c066e731ff64c265d886896a9) — fix: append \u002Fmcp suffix to Cloud URL to match C# Host property by @IvanMurzak\n- [`86e10ca0`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F86e10ca01508bd0e5028e3d044a43faae52c9718) — fix: serialize enums as strings in config JSON, handle legacy integers by @IvanMurzak\n- [`3f4338be`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F3f4338be7511a0e2cb94939372111c0342f82260) — feat: auto-resolve URL and token from project config in run-tool by @IvanMurzak\n- [`5a283d82`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F5a283d82b8b8c9109b45f7d9d505217f455b5d12) — fix: correct namespace import in UnitySkillFileGenerator by @IvanMurzak\n- [`ad3be8e4`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fad3be8e4e55f5226bda2c00cc7b6ba255f34ffd7) — feat: override skill generation to use unity-mcp-cli run-tool syntax by @IvanMurzak\n- [`45da66b5`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F45da66b55a251c0cfe0edbeb9beb6cac36b033ae) — feat: add run-tool CLI command for direct MCP tool execution by @IvanMurzak\n- [`ab4a39a3`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fab4a39a3c3424feebf65204","2026-03-16T08:10:24",{"id":267,"version":268,"summary_zh":269,"released_at":270},71047,"0.54.0","# AI Game Developer (Unity MCP) 0.54.0\n**Released:** *March 15, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.53.1...0.54.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.53.1...0.54.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`b7e78f3f`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fb7e78f3fbaed713fd72a897059f26ecdb68eb2dd) — Bump component versions to 0.54.0 by @IvanMurzak\n- [`af3d38b9`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Faf3d38b99999cd60077b06ff229068a4ab6c955d) — feat: refactor CLI — merge open+connect, TTY detection, verbose flag (#572) by @IvanMurzak\n- [`320755eb`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F320755ebcd7b43efecad958f0f0d69314a5b5e21) — Add console-clear-logs tool (#465) (#562) by @cagilfirat\n- [`590c9b6f`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F590c9b6f53ac2035374e2fb1f2792514b2b90dfc) — Add mandatory MCP tool arg and docs sync rules by @IvanMurzak\n- [`93b4a49f`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F93b4a49fb6cd00de3d6cd5014beb86ec00a675d8) — Prohibit reflection for non-public C# access by @IvanMurzak\n- [`be690b1d`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fbe690b1db1f6f6731ed14e677dfab4e54655f155) — Add publish-test-results workflow & upload event (#571) by @IvanMurzak\n- [`43dc51d0`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F43dc51d02afaab0434c70585296a24dc0ae648bc) — fix: replace --version option with positional argument in install-unity command (#570) by @IvanMurzak\n- [`c4c74881`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc4c74881737f985fef0ca24961882b6a79303d41) — Feature\u002Fadd cline configurator (#565) by @mainVar\n- [`f5545d02`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Ff5545d02140ad969177b6450403834d8a519e5f2) — Refresh agent configs and reconnect after cloud authorization (#569) by @IvanMurzak\n- [`17681bf3`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F17681bf35c57943f0ecbfdeb0f16b727b911f0df) — Add speckit integration and optimize CLAUDE.md docs (#568) by @IvanMurzak\n- [`d71853c2`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fd71853c29e8920bfef2def4478fba630359b857c) — Refactor CLI commands and update deps (#567) by @IvanMurzak","2026-03-15T02:55:23",{"id":272,"version":273,"summary_zh":274,"released_at":275},71048,"0.53.1","# AI Game Developer (Unity MCP) 0.53.1\n**Released:** *March 14, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.53.0...0.53.1](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.53.0...0.53.1)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`5ca69ab3`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F5ca69ab3451989823435da870a8df98ad25f981d) — Bump version to 0.53.1 by @IvanMurzak\n- [`c4afcdc0`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc4afcdc08ff7ec7ba76d381f9e2e70dd273aa6ad) — docs: fix outdated --start-server flag syntax in CLI docs by @IvanMurzak\n- [`640a59a5`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F640a59a53014e5f740819e38b40d8eb3cae690a0) — fix: add --allow-same-version to npm version in deploy workflow by @IvanMurzak","2026-03-14T04:22:28",{"id":277,"version":278,"summary_zh":279,"released_at":280},71049,"0.53.0","# AI Game Developer (Unity MCP) 0.53.0\n**Released:** *March 14, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.52.0...0.53.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.52.0...0.53.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`98d29868`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F98d2986895c1e96edb3a1b4160cf063d4c90992c) — Update UnityMcpPlugin.Config.cs by @IvanMurzak\n- [`590f2b36`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F590f2b3690e6002f176bf5229cf3402b356d9946) — Add resolvedVersion overload and expand tests by @IvanMurzak\n- [`0fe7066a`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F0fe7066ad460e684886334e223b2e04b7734e942) — Update test workflow to publish test results and refine version comparison assertion by @IvanMurzak\n- [`57a2ed14`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F57a2ed14c7ab2ce289134dc2ed70269fd455a80c) — Bump version to 0.53.0 across all components and configuration files by @IvanMurzak\n- [`0cbb1632`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F0cbb16324d63a4b2ee7507af84f75a798f368ce9) — Update default connection settings to keep connection and server running by @IvanMurzak\n- [`833f9601`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F833f9601c938b1b908ffc863f118248f1b0b2c61) — Disable various prompts and tools in the MCP Plugin (#566) by @IvanMurzak\n- [`fe54f2fa`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Ffe54f2fa590807d4f2c1683594ee3d6ece828423) — Add fancy terminal UI with chalk, ora, and boxen to CLI (#564) by @IvanMurzak\n- [`e99e688d`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fe99e688d39287791d95232b20eae2a159300e48e) — Add remove-plugin command to uninstall Unity-MCP from projects (#561) by @IvanMurzak\n- [`ded78ca5`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fded78ca5aed4243dc01470769bf8dbe00fe0facc) — Update deployment workflow: remove auto-release trigger and improve NuGet auth (#560) by @IvanMurzak\n- [`305cb470`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F305cb470f14c0229bf5eeebde2cd0b202eb98deb) — Add OpenUPM version resolution to prevent installation failures (#559) by @IvanMurzak\n- [`b234034e`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fb234034ea8d55659145916c5a473f709ff9c8c0d) — Fix Node.js 20 deprecation warnings in release.yml (#558) by @IvanMurzak","2026-03-14T03:47:17",{"id":282,"version":283,"summary_zh":284,"released_at":285},71050,"0.52.0","# AI Game Developer (Unity MCP) 0.52.0\n**Released:** *March 13, 2026*\n\n---\n\n## Comparison\nSee every change: [Compare 0.51.6...0.52.0](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcompare\u002F0.51.6...0.52.0)\n\n---\n\n## Commit Summary (Newest → Oldest)\n- [`6b2f39bd`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F6b2f39bdb1804ac34b467b15846e35553287efad) — Merge branch 'main' of https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP by @IvanMurzak\n- [`c09136d4`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fc09136d4a02b721fb5d9d534363cf1b82976f445) — chore: update version to 0.52.0 in multiple files and documentation by @IvanMurzak\n- [`1f27114a`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F1f27114af5bbc0fcacce8e90d84244d5a4273e6f) — ci: upgrade GitHub Actions to Node.js 24-compatible versions (#557) by @IvanMurzak\n- [`3856a609`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F3856a60963b342e6de91ca20b0d6c22b504c7258) — Restrict Claude Code workflow to only trigger for IvanMurzak (#556) by @IvanMurzak\n- [`0b1be141`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F0b1be1417112db8b00975494a93c9807a035a937) — Update plugin DLL binaries (#555) by @IvanMurzak\n- [`63a35ffe`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F63a35ffeaa5b972a1462314d867cb64a343e81de) — Fix dual-stack socket crash on macOS (#553) by @IvanMurzak\n- [`79501da4`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F79501da43397f11fdb33a109bda8674051aa5145) — feat: installer self-deletes after successful installation (#554) by @IvanMurzak\n- [`421a6f71`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F421a6f712bd8f91b3ae6b24ebbaeede2fcf9f7d9) — fix: add dummy parameter to zero-arg MCP tools for Copilot compatibility (#552) by @IvanMurzak\n- [`152f1f4d`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F152f1f4db48d8e19c9474e25308c100f1b64e29d) — Update ai-connector-window.gif by @IvanMurzak\n- [`19cc3a88`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F19cc3a88778e150b8e4ea2d90e825e1a05903d15) — Add cross-platform npm CLI tool for Unity-MCP operations (#493) (#549) by @IvanMurzak\n- [`3345b324`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F3345b324ffbba78511cdf5be40930e94f1c95c45) — Setup NuGet trusted publishing via OIDC to replace expiring API key (#551) by @IvanMurzak\n- [`d97af103`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002Fd97af1034eed17713c836eaa129c623e0187a835) — Enhance cloud device authentication and connection management features (#550) by @IvanMurzak\n- [`3609ccce`](https:\u002F\u002Fgithub.com\u002FIvanMurzak\u002FUnity-MCP\u002Fcommit\u002F3609cccee2474170a969e38e9a74e246b78ec38d) — Implement cloud device authentication and enhance MCP server management (#548) by @IvanMurzak","2026-03-13T08:28:56"]