chathook
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@chathook 帮我写一封邮件,然后询问是否需要修改"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
chathook — AI 对话钩子 MCP Server
在 AI 完成任务后弹出原生输入对话框,让用户直接输入回复,保持对话不中断,直到用户明确结束。
工作原理
用户提问 → AI 处理任务 → AI 调用 chat_hook 工具
↓
弹出原生桌面对话框(输入框 + 提交按钮 + 倒计时)
↓
用户输入回复 / 关闭对话框结束
↓
工具返回用户输入 → AI 继续处理
↓
AI 再次调用 chat_hook ...
↓
用户关闭对话框 → 工具返回结束信号
↓
AI 结束回合,对话完成Related MCP server: mcp-interactive-ui-server
安装
方式一:npx 从 npm 安装(推荐)
在 MCP 客户端的配置中添加:
{
"mcpServers": {
"chathook": {
"command": "npx",
"args": ["-y", "mcp-chathook"]
}
}
}方式二:npx 从 GitHub 安装
{
"mcpServers": {
"chathook": {
"command": "npx",
"args": ["-y", "github:kedaya2025/mcp-chathook"]
}
}
}方式三:本地运行
git clone https://github.com/kedaya2025/mcp-chathook.git
cd mcp-chathook
npm install
npm run build{
"mcpServers": {
"chathook": {
"command": "node",
"args": ["<path-to>/mcp-chathook/dist/index.js"]
}
}
}使用方法
系统提示词配置
在 AI Agent 的系统提示词 / 自定义指令中添加:
所有对话或任务结束时,都必须调用 chathook 工具来向用户确认工作进程,
除非用户结束该工具的调用视为任务完成不再调用。License
MIT
Available Tools
1 toolchat_hookA
Display an input dialog to the user and wait for their response.
IMPORTANT: After completing a task or response, you SHOULD call this tool to ask the user if they need anything else. This keeps the conversation alive without requiring the user to type a new message.
The tool will block until the user responds. Possible return values:
Normal text: The user's new request or feedback. Continue working.
[CHATHOOK:USER_DECLINED]: The user clicked 'Decline' (end conversation). Stop calling this tool and end your response.
[CHATHOOK:USER_CANCELLED]: The user cancelled the input. You may end your response or call this tool again.
Inputs: message (required), suggestions.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 1 tool
With only one tool, there is no possibility of confusion. The tool's purpose is clearly distinct.
The single tool name 'chat_hook' uses a consistent underscore convention, which is fine for a single tool.
The server has only one tool, which is at the low end of acceptable. While it serves a specific purpose, typically servers have more tools to cover broader functionality.
The single tool fully covers the server's purpose of user interaction. It handles input and special cases (decline, cancel) through return values, so no gaps are apparent.
Maintenance
Related MCP Connectors
An AI concierge that turns static forms into adaptive AI conversations. From any MCP client.
Let your AI agent ask a person for a decision, a date or a photo, and get a structured answer back.
An MCP connector that lets your AI agents ask you questions, request permission approvals, and send alerts as push notifications to your phone and mac notch, so you can unblock and approve your agents from anywhere.
Turn your offers into an AI agent. Set it up from your MCP assistant, answer customer questions, collect enquiries and connect with other agents via A2A. OAuth sign-in; free plan available.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables AI models to interactively prompt users for input or clarification directly through their code editor. It facilitates real-time communication between assistants and users during development tasks.MIT
- AlicenseAqualityDmaintenanceEnables AI agents to render interactive user interfaces such as forms, dashboards, charts, tables, and wizards directly in MCP-compatible clients. Supports structured data collection and richer interactions beyond text responses.613 PyPIMIT
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to ask clarification questions and receive structured user input through a Human-in-the-Loop interface.1MIT
- AlicenseAqualityAmaintenanceAn MCP server that enables human-in-the-loop elicitation, letting AI agents ask users questions via tools like elicit_confirm and elicit_form, with diagnostics to verify host elicitation support.41Apache 2.0