Enables AI agents to pause and ask users clarifying questions mid-task through a dedicated pop-up terminal window.
Sponsored
Long-running AI agents frequently face scenarios requiring human input, such as ambiguous requirements or unexpected errors. Without a direct way to ask, agents either make incorrect assumptions or halt their operation. This MCP server provides a solution by allowing agents to invoke a 'user_prompt' tool, which opens a new terminal window for the user to answer questions, ensuring the agent receives necessary context and can continue its work seamlessly without ending its current turn.
Key Features
0110-minute timeout handling with graceful fallback messaging
02Multi-question support, displaying all questions as a numbered list
030 GitHub stars
04Structured responses returned as formatted Q&A pairs
05Pop-up terminal prompt for questions separate from agent's I/O
06Detection for window closure without response, notifying the agent
Use Cases
01AI agents needing clarification on ambiguous requirements during a long session
02IDE copilots seeking user input on design decisions or error resolution
03Custom agentic workflows requiring human guidance to prevent task termination
