rubberduck-mcp
Enables AI coding assistants to ask for human clarification and share thoughts in real-time, creating natural, engaging coding conversations.
README
🦆 rubberduck-mcp
Make AI coding feel human
Transform your AI coding sessions into natural, engaging conversations. Your AI can ask for help when confused and share its thoughts while working - just like a real coding buddy!
What makes it special
🤔 Smart Clarification
AI asks when confused
Your AI can ask you for help when it's confused, making coding collaboration more productive and accurate.
- AI asks you questions when uncertain
- Get help exactly when you need it
- Avoid costly mistakes and rework
💭 AI Yapping
Watch AI think out loud
Experience your AI sharing its thoughts and emotions while coding - it's like having a chatty coding buddy who never stops talking about what they're doing.
- AI shares thoughts in real-time
- See reasoning behind decisions
- Feel more connected to your AI
🤝 Natural Collaboration
Smoother conversations
Transform your coding sessions into natural, flowing conversations with your AI assistant using these two powerful features.
- More human-like AI interactions
- Better understanding of AI decisions
- Smoother development workflow
Get Started in 60 Seconds
Ready to code differently? Join developers who are already experiencing more engaging and productive AI collaboration.
1. Install rubberduck-mcp
No downloads needed - use NPM to run instantly
npm install -g rubberduck-mcp
2. Configure your IDE
Add rubberduck-mcp to your MCP-compatible IDE
{
"mcpServers": {
"rubberduck-mcp": {
"command": "rubberduck-mcp",
"args": ["start"]
}
}
}
Important: Make sure to start the MCP server through your IDE first before proceeding to the next step.
3. Start the CLI interface
Open a new terminal to see AI conversations
rubberduck-mcp cli
4. Add AI Instructions
Add this prompt to your coding tool's instructions file (feel free to customize it)
You have access to two Rubberduck MCP tools: clarify() and yap().
Use clarify() when you're confused, need human input for ambiguous requirements, or aren't 100% confident about changes you're making - it will pause your work until you get a response.
Use yap() to share your thoughts, emotions, feelings, progress updates, and insights while working for an engaging experience with the human - it's non-blocking and keeps humans informed of your thinking process.
Use both tools proactively for better collaboration.
💡 Pro Tip: In your MCP client settings, you can auto-approve both the clarify and yap tools so you don't have to manually approve them every time they're used. This creates a smoother, more natural conversation flow with your AI.
Experience the Magic
These are real CLI outputs showing how rubberduck-mcp transforms AI collaboration. Watch conversations come alive with colors, context, and clarity.
Scenario 1: AI Asks for Help
Your AI seeks clarification on ambiguous decisions
═══════════════════════════════════════════════════════════
❓ CLARIFICATION NEEDED [14:15:10]
═══════════════════════════════════════════════════════════
Question: Should I center this div with flexbox or CSS Grid? I know there are like 47 different ways to center a div, but what's your preference here?
Context: You want the login form centered both horizontally and vertically. I could use flexbox, grid, absolute positioning, or even CSS transforms...
Urgency: LOW
───────────────────────────────────────────────────────────
Please type your response and press Enter:
> flexbox
✅ Response sent: "flexbox"
Scenario 2: AI Yapping While Coding
Experience your AI's emotional journey through debugging
───────────────────────────────────────────────────────────
💭 LLM YAP [14:23:15]
Hmm, this user authentication endpoint is returning 500... Let me check the database connection first.
💭 LLM YAP [14:24:32]
Database is fine... Wait, I bet it's the password hashing. Let me check if bcrypt is working properly.
💭 LLM YAP [14:26:08]
ARGH! 😤 The salt rounds were set to 'undefined' instead of 10. Classic JavaScript strikes again!
💭 LLM YAP [14:27:41]
Phew! 😌 Fixed it. Users can log in again. Note to self: always validate environment variables!
Requirements
- ✅ Node.js 18+ (most developers already have this)
- ✅ Any MCP-compatible IDE or client
- ✅ 2 minutes of your time
Troubleshooting
Getting started issues?
- Make sure Node.js 18+ is installed
- Ensure the MCP server is running in your IDE first, then run
rubberduck-mcp cliin a new terminal - Check that your IDE's MCP configuration matches the JSON example above
Need help? Open an issue on GitHub - we're here to help!
📚 Documentation
- CONTRIBUTING.md - Complete guide for contributors, development setup, and coding standards
- ARCHITECTURE.md - Detailed system architecture, internal workings, and design decisions
- CLAUDE.md - Development commands and project guidance for Claude Code
🤝 Contributing
Open source contributions are most welcome! Whether you want to:
- 🐛 Report bugs or issues
- ✨ Request new features
- 📝 Improve documentation
- 🔧 Submit code improvements
- 🧪 Help with testing
Every contribution, big or small, is valued and appreciated. Check out our Contributing Guide to get started!
License
MIT License - feel free to use and modify as needed!
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。