User Feedback

User Feedback

Simple MCP Server to enable a human-in-the-loop workflow in tools like Cline and Cursor.

50
Stars
10
Forks
0
Releases

Overview

User Feedback MCP is a lightweight MCP server designed to enable human-in-the-loop workflows within MCP-enabled tools such as Cline and Cursor. It provides a user interface to collect user feedback during interactive tasks and supports testing of desktop applications that require complex user interactions. The README outlines installation with uv, registering the server in Cline's MCP Servers configuration, and a sample .user-feedback.json that controls how commands are run and whether they execute automatically on startup. A recommended prompt snippet guides LLM prompts to invoke the user_feedback MCP tool before task completion, ensuring feedback is gathered before marking work as complete. The project exposes a web interface at http://localhost:5173 for testing MCP tools and demonstrates an Available tools block that shows how to call the user_feedback tool with a JSON argument object. This setup emphasizes testability and repeatable steps for integrating user feedback into AI-assisted workflows.

Details

Owner
mrexodia
Language
Python
License
MIT License
Updated
2025-12-07

Features

Human-in-the-loop workflow

Provides an MCP server that enables asking for user feedback before completing tasks in MCP-enabled tools like Cline and Cursor.

Prompt engineering guidance

Includes a recommended prompt snippet to force the user_feedback MCP tool to solicit user feedback before task completion.

.user-feedback.json configuration

Saves a .user-feedback.json with command and auto-run behavior; controls automated execution on startup.

Cline integration guide

Step-by-step installation and registration in Cline's MCP Servers configuration to run the server.

Local development web UI

Runs via uv with a web interface at http://localhost:5173 for testing MCP tools.

Tool invocation example

Demonstrates the Available tools block showing how to call the user_feedback tool with arguments.

Audience

DevelopersEnable human-in-the-loop feedback during interactive testing in MCP-enabled tools for desktop apps.
QA EngineersCollect user feedback during interactive testing of desktop apps in MCP environments.
Product teamsValidate task flows with user feedback before final releases in MCP workflows.

Tags

mcphuman-in-the-loopuser-feedbackclinecursordesktop-app-testingprompt-engineeringautomationweb-uiuvserver