Human-in-the-Loop Slack MCP Server

A Model Context Protocol (MCP) server that enables AI assistants to request information from humans via Slack. This server acts as a bridge between AI systems and human experts, allowing AI to ask questions and receive responses through Slack when it needs human knowledge or clarification.

Quick Start with npx

Run directly from GitHub without installation:

Example with Claude Desktop

Add to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json):

Features

• 🤖 MCP-compliant server for AI assistant integration

• 💬 Real-time Slack integration via Socket Mode WebSocket connection

• 🧵 Thread-based conversations for maintaining context

• ⏱️ 60-second timeout for human responses

• 📢 User mentions (@username) for notifications

• 🔍 Comprehensive debugging and logging capabilities

• 🔐 Secure token handling

• 🚀 Dynamic handler initialization for faster startup

• ⚡ Optimized for instant response detection with event-driven architecture

Prerequisites

1. Slack App Setup

   • Create a new Slack app at https://api.slack.com/apps

   • Enable Socket Mode in your app settings

   • Generate an App-Level Token with connections:write scope

   • Install the app to your workspace

2. Bot Token Scopes

   • chat:write - Send messages

   • channels:read - Access channel information

   • users:read - Access user information

3. Socket Mode

   • Enable Socket Mode in your app settings

   • This is required for the app to receive events in real-time

4. Event Subscriptions

   • Enable Events API

   • Subscribe to bot events:

     • message.channels - Messages in public channels

     • message.groups - Messages in private channels

     • message.im - Direct messages (optional)

   • Save changes and reinstall the app to your workspace

Installation (Optional)

If you want to install locally instead of using npx:

1. Clone the repository:

2. Install dependencies:

3. Build the TypeScript code:

Configuration

All configuration is passed via command-line arguments:

• --slack-bot-token - Bot User OAuth Token (xoxb-...)

• --slack-app-token - App-Level Token for Socket Mode (xapp-...)

• --slack-channel-id - Channel ID where the bot will operate

• --slack-user-id - User ID to mention when asking questions

• --log-level - (Optional) Logging level (default: INFO)

Usage

Development Mode

Run with hot-reloading:

Production Mode

Build and run:

With MCP Client (Using npx)

Configure your MCP client to use this server directly from GitHub:

With MCP Client (Local Installation)

If you've installed locally:

Available Tools

ask_on_slack

Main tool for asking questions to humans via Slack.

Parameters:

• question (string): The question to ask the human. Be specific and provide context.

Example:

Usage Notes:

• The bot will mention the specified user in the Slack channel

• The human has 60 seconds to respond in a thread

• The tool will return the human's response or timeout after 60 seconds

Development

Scripts

• npm run build - Compile TypeScript

• npm run dev - Run with hot-reloading

• npm start - Run compiled code

• npm test - Run tests with Vitest

• npm run test:ci - Run tests with coverage

• npm run lint - Run ESLint

• npm run format - Format code with Prettier

• npm run clean - Clean build artifacts

Project Structure

Testing

The project uses Vitest for testing. Tests are located in the tests/ directory.

To run tests:

CI/CD

The project uses GitHub Actions for continuous integration and deployment.

• CI Workflow (ci.yml): Runs on every push and pull request

  • Tests on Node.js 18.x, 20.x, and 22.x

  • Runs linting and type checking

  • Generates code coverage reports

  • Builds the project

• Release Workflow (release.yml): Runs on version tags

  • Builds and tests the project

  • Creates GitHub releases

  • Publishes to npm (requires NPM_TOKEN secret)

Troubleshooting

1. Connection Issues

   • Verify all tokens are correct

   • Check that the bot is invited to the channel

   • Ensure Socket Mode is enabled in your Slack app

2. No Response Received

   • Verify the user ID is correct (format: U1234567890)

   • Ensure the user responds in the message thread, not the main channel

   • Check that the bot has permission to read messages in the channel

3. Authentication Errors

   • Bot token should start with xoxb-

   • App token should start with xapp-

   • Regenerate tokens if needed

   • Verify bot has required scopes: chat:write, channels:read, users:read

4. Performance Optimization

   • The server uses event-driven architecture for instant response detection

   • WebSocket connection ensures real-time message delivery

   • Detailed timing logs available with [TIMING] prefix for debugging

License

MIT