User Feedback MCP

What is User Feedback MCP?

User Feedback MCP is a simple MCP server that introduces a human-in-the-loop workflow into AI‑assisted coding tools like Cline and Cursor. It is designed for developers who need to collect user feedback before completing tasks, especially when testing complex desktop application interactions.

How to use User Feedback MCP?

Install the uv package manager globally, clone this repository, then add the server to your Cline MCP configuration file (cline_mcp_settings.json) using the command uv --directory <path> run server.py. Optionally add the custom prompt “Before completing the task, use the user_feedback MCP tool to ask the user for feedback.” to your prompt. The server reads a .user-feedback.json file in your project directory to optionally auto‑execute a command.

Key features of User Feedback MCP

• Human‑in‑the‑loop feedback for AI coding assistants.
• Works with Cline, Cursor, and other MCP‑compatible tools.
• Configurable command execution via .user-feedback.json.
• Simple installation using uv and a cloned repository.
• Provides a single user_feedback tool with project directory and summary arguments.
• Web UI for testing during development (via uv run fastmcp dev server.py).

Use cases of User Feedback MCP

• Ask for user approval before marking a code‑edit task as completed.
• Test desktop applications that require manual interaction steps.
• Validate complex changes by presenting a summary to the user.

FAQ from User Feedback MCP

What is the purpose of this MCP server?

It enables a human‑in‑the‑loop workflow, allowing AI coding tools to request feedback from the user before finalizing a task.

How do I install User Feedback MCP?

Install uv globally (e.g., pip install uv on Windows or the official install script on Linux/Mac), clone the repository, then add the server configuration to your Cline MCP settings.

What configuration file does it use?

The server reads a .user-feedback.json file in your project directory. It can contain a command to run and an execute_automatically flag to run it without clicking “Run”.

What tools does User Feedback MCP expose?

It exposes one tool called user_feedback. The tool requires a project_directory and a summary argument.

Does it have any runtime dependencies?

Yes, it requires Python and the uv package manager. The server is started using uv run server.py.