interactive-feedback-mcp

# Interactive Feedback MCP - Interactive Feedback Collector > Based on the original project by Fábio Ferreira ([@fabiomlferreira](https://x.com/fabiomlferreira)) > Reference project: [sanshao85/mcp-feedback-collector](https://github.com/sanshao85/mcp-feedback-collector) > Current maintainer: bulice > For more AI development enhancement tools, please visit [dotcursorrules.com](https://dotcursorrules.com/) ## 📖 Project Overview Interactive Feedback MCP is a server based on [MCP (Model Context Protocol)](https://modelcontextprotocol.io/) designed for human-computer interaction feedback systems tailored for AI-assisted development tools such as [Cursor](https://www.cursor.com), [Cline](https://cline.bot), and [Windsurf](https://windsurf.com). This project is a secondary development and optimization based on the original design by Fábio Ferreira, while also referencing the modern implementation of the [sanshao85/mcp-feedback-collector](https://github.com/sanshao85/mcp-feedback-collector) project. By providing a graphical interface, it allows users to give real-time text feedback, upload images, execute commands, and view outputs during the AI development process, thus achieving a true human-computer collaborative development model. ## ✨ Core Features ### 🎯 Interactive Feedback - **📝 Text Feedback**: Provide detailed text feedback to the AI assistant - **🖼️ Image Support**: Support for multiple image uploads or clipboard pasting - **💬 Real-time Interaction**: Engage in real-time conversation and feedback with the AI assistant ### 🎨 User Interface - **🌙 Dark Theme**: Eye-friendly dark interface, suitable for long-term use - **☀️ Light Theme**: Refreshing light interface, suitable for bright environments - **🎛️ Theme Switching**: Supports dynamic switching of interface themes - **📱 Responsive Design**: Adapts to window size and layout ### ⚙️ Command Execution - **🔄 Real-time Output**: Displays the command execution results in real-time - **📊 Process Monitoring**: Monitors the command execution status and process information - **🚀 Automatic Execution**: Optionally execute commands automatically at startup - **💾 Command History**: Saves and manages frequently used commands ### 🗂️ Project Management - **📁 Project-Specific Settings**: Each project saves configurations independently - **⚙️ Configuration Persistence**: User preferences are saved using Qt QSettings - **🔧 Flexible Configuration**: Supports various startup and configuration methods ## 🏗️ Technical Architecture ### Core Technology Stack - **Python 3.11+**: Main development language - **FastMCP**: MCP server framework - **PySide6**: Cross-platform GUI framework - **psutil**: System process management - **Pillow**: Image processing library ### Project Structure ``` interactive-feedback-mcp/ ├── server.py # MCP server main program ├── feedback_ui.py # GUI interface implementation ├── diagnose_mcp.py # MCP connection diagnostic tool ├── test_mcp.py # MCP server test script ├── mcp_server.sh # Server startup script ├── requirements.txt # Python dependencies ├── pyproject.toml # Project configuration file ├── cursor_mcp_config.json # Cursor configuration example ├── images/ # Image resources directory ├── logs/ # Log files directory └── README.md # Project documentation ``` ### Core Components #### 1. MCP Server (`server.py`) - Implements the server side of the MCP protocol - Provides `interactive_feedback` and `get_image_info` tools - Handles communication with the AI assistant #### 2. GUI Interface (`feedback_ui.py`) - Modern interface based on PySide6 - Supports dark/light theme switching - Implements functions such as text input, image upload, and command execution #### 3. Diagnostic Tool (`diagnose_mcp.py`) - Check the connection status of the MCP server - Validate dependencies and configurations - Generate configuration recommendations ## 🚀 Installation and Configuration ### System Requirements - **Python**: 3.11 or higher - **Operating System**: Windows, macOS, Linux - **Package Manager**: [uv](https://github.com/astral-sh/uv) (recommended) or pip ### Installation Steps #### 1. Get the Code ```bash git clone https://github.com/bulice/interactive-feedback-mcp.git cd interactive-feedback-mcp ``` #### 2. Install Dependencies ```bash ``` # Using uv (Recommended) uv sync # Or use pip ```bash pip install -r requirements.txt ``` #### 3. Running the Server ```bash # Using uv uv run server.py # Or run directly ```bash python server.py ``` ### Cursor Configuration Add the following configuration to the MCP configuration file of Cursor: ```json "mcpServers": { "interactive-feedback-mcp": { "command": "/Users/lee/tmp/interactive-feedback-mcp/mcp_server.sh", "args": [], "timeout": 600, "autoApprove": [ "interactive_feedback" ] } } ``` **Note**: Please replace `/path/to/interactive-feedback-mcp` with the actual project path. ### Other AI Tool Configuration For tools like Cline and Windsurf, the configuration method is similar; you just need to specify the server commands and parameters in the corresponding MCP settings. ## 📋 Usage Instructions ### Basic Usage Process 1. **Start the Server**: Run the MCP Server 2. **Configure AI Tools**: Configure the MCP Server in tools like Cursor 3. **AI Invocation**: The AI assistant invokes the feedback tool via the MCP protocol 4. **User Interaction**: Provide feedback in the pop-up interface 5. **Feedback Transmission**: Feedback information is returned to the AI assistant ### Prompt Engineering To achieve the best results, it is recommended to include the following rules in the custom prompts for the AI assistant: ``` Always invoke the MCP `interactive_feedback` tool when you want to ask a question. When you are about to complete the user request, call the MCP `interactive_feedback` tool instead of directly ending the process. Continue to invoke MCP until the user feedback is empty, then end the request. ``` ### Topic Selection ```bash ``` # Using Dark Theme (Default) python feedback_ui.py --theme dark --prompt "Your message" # Using Light Theme ```bash python feedback_ui.py --theme light --prompt "Your message" ``` ### Command Line Arguments - `--project-directory`: Specify the project directory - `--prompt`: Set the prompt message - `--theme`: Choose the interface theme (light/dark) - `--output-file`: Specify the output file path ## 🔧 Development Debugging ### Development Mode Run ```bash uv run fastmcp dev server.py ``` This will start a web interface for testing the MCP tools. ### Connection Diagnosis ```bash python diagnose_mcp.py ``` Run the diagnosis tool to check the connection status of the MCP server. ### Test Server ```bash python test_mcp.py ``` Test the basic functionality of the MCP Server. ## 💡 Value Proposition By guiding the AI assistant to confirm with the user before completing tasks, rather than making speculative high-cost tool calls, this module can significantly reduce the number of advanced requests on platforms like Cursor. In some cases, it can consolidate operations that would originally require 25 tool calls into a single feedback-aware request, thereby saving resources and improving performance. ## 📁 Configuration Management The project uses Qt's `QSettings` to store configurations on a per-project basis, including: - The command to be executed - Whether to automatically execute the command on the next startup - The display/hide status of the command area - The window geometry and state These settings are typically stored in platform-specific locations (Windows Registry, macOS plist files, Linux configuration files, etc.). ## 🤝 Acknowledgments and Contact Information ### Acknowledgments for the Original Projects This project is developed based on the following excellent projects: - **Original Idea**: Fábio Ferreira's [Interactive Feedback MCP](https://x.com/fabiomlferreira) - **Modern Implementation**: [sanshao85/mcp-feedback-collector](https://github.com/sanshao85/mcp-feedback-collector) Thanks to these developers for their original contributions and open-source spirit. If you find this project useful, we encourage you to follow the work of the original authors. ### Current Maintenance - **Maintainer**: bulice - **Project URL**: https://github.com/bulice/interactive-feedback-mcp - **Issue Feedback**: Please submit issues on the GitHub project page ### Related Resources Please visit [dotcursorrules.com](https://dotcursorrules.com/) for more enhanced resources on AI-assisted development workflows. ## 📄 License This project is licensed under an open-source license. Please refer to the [LICENSE](LICENSE) file for details.