"Allowing an AI agent to ask user questions during a task" matching MCP servers:

Enables real-time user intervention for MCP agents through a web UI, allowing users to review context and provide feedback when AI agents drift from intent, keeping them on track.

Enables AI-assisted user feedback collection and management via a web interface, with MCP integration for automated replies and tool execution.

Provides A-share (Chinese stock market) quantitative analysis through tools for stock screening, northbound capital flow tracking, dragon-tiger list analysis, margin trading, sector analysis, technical indicators, IPO info, and limit-up/down statistics using akshare data.

A multi-modal user intent collection MCP server that allows AI agents to ask users questions and receive text/image replies.

Converts natural language queries into valid GraphQL queries and executes them against GraphQL APIs. Includes schema introspection, query validation, execution with authentication, and query history tracking.

Enables searching and retrieving documentation from crawled documentation sites as an MCP server, allowing coding agents to query real docs instead of relying on training data.

Enables conversion of YouTube videos to MP3 format through the Youtube To Mp315 API. Supports checking conversion status, retrieving video titles, and asynchronous video-to-audio conversion with customizable quality and time range settings.

A lightweight MCP bridge that lets Claude Code delegate heavy tasks to the Antigravity CLI - saving context window and tokens for what matters.

Enables AI assistants to manage tasks with tools for adding, completing, and deleting tasks, and resources for viewing all or pending tasks.

Fetches a URL and returns the main content as clean Markdown, using plain HTTP when possible and headless Chromium for JavaScript-rendered or bot-protected pages.

A self-improving knowledge agent that provides local retrieval, web research, and structured answer synthesis for Claude Code and VS Code Copilot via MCP. It enables AI tools to manage a project-specific knowledge lifecycle through automated research and indexing.

MCP server that lets you drive Claude Code sessions hands-free by voice from any MCP client, enabling remote models to send prompts, monitor progress, and get results read back.

A free, read-only server that lets users diagnose production AI-agent risks: it flags over-privileged API keys from permission descriptions, scans code snippets for the signals that prevent duplicate execution, and explains guard patterns such as idempotency keys, approval gates, outboxes, and circuit breakers. It makes no network calls or writes, and rejects inputs that look like real secrets.

Connects MCP clients directly to a live Godot editor, enabling AI assistants to build scenes, edit nodes, and control the editor through over 120 operations.

Universal session & handoff layer for AI coding agents: shared task, decision, blocker and test state, checkpoints, and deterministic handoff between Claude Code, Codex, OpenCode, Cursor, Cline, pi and omp. Local SQLite, git-aware, 25 tools + 7 resources over stdio or streamable-http.

A Model Context Protocol (MCP) server that exposes the EasyDeploy public REST API as tools for Claude, Cursor, Claude Code, and other MCP clients.

An MCP server that acts as a proxy to connect MCP clients to agent frameworks like Mastra and LangGraph, enabling agent discovery, dynamic server connections, and recursive agent networks.

Official Bitget MCP server for Claude Desktop, Cursor, Windsurf & ChatGPT. Trade crypto via natural language — 89 UTA v3 ops through 14 intent verbs. Local stdio, HMAC signing, read-only & paper-trading modes.

MCP server for AI-driven VAPT orchestration, enabling agents to plan and execute authorized security scans through a control plane that enforces scope, sanitization, budget, rate limits, human approval, and audit logging.

13 MCP tools that plan and mechanically validate color systems for front-end work: OKLCH role derivation, WCAG + APCA contrast gates, color-blind simulation, AI-slop detection, 10-scene visualization and one-verdict reporting over a 117-palette evidence library.