Feedback Loop

Provides a human-in-the-loop feedback collection system for AI-assisted development tools like Cursor, Cline, and Windsurf through an Electron-based GUI that launches on-demand to gather user input. Built by Tuan Dinh using Electron and the Model Context Protocol SDK, the implementation features a single feedback_loop tool that spawns a draggable, always-on-top window with customizable quick feedback options, free-form text input, and keyboard shortcuts for efficient interaction. The system captures user feedback through a glass-effect interface positioned strategically on screen, returns structured JSON responses with timestamps and project context, and automatically handles window lifecycle management including proper cleanup on cancellation, making it valuable for development workflows requiring human oversight of AI-generated code changes, interactive debugging sessions where developer input guides AI decisions, and collaborative coding environments that benefit from real-time human feedback integration.