Get the FREE Ultimate OpenClaw Setup Guide →

mcp-notify

Minimal MCP server for sending Slack messages as user — with bot signature and concurrent safety

Installation
Run this command in your terminal to add the MCP server to Claude Code.
Run in terminal:
Command
claude mcp add --transport stdio pilat-mcp-notify npx -y @pilat/mcp-notify \
  --env SLACK_MCP_DATA_DIR="~/.local/share/mcp-notify" \
  --env SLACK_MCP_BOT_TOKEN="xoxb-..." \
  --env SLACK_MCP_XOXC_TOKEN="xoxc-..." \
  --env SLACK_MCP_XOXD_TOKEN="xoxd-..."

How to use

mcp-notify is a minimal MCP server that sends Slack messages using a single tool called send_message. It supports two authentication modes: bot token mode (xoxb) and user-token mode (xoxc + cookie). In bot-token mode, messages are sent as the Slack app with appropriate branding; in user-token mode, each message includes a :robot_face: context block to indicate AI-assisted origin. The server uses Block Kit with a plain-text fallback and stores channel/user data in a SQLite cache to improve performance and enable concurrent access across multiple MCP instances.

How to install

Prerequisites:

  • Node.js and npm installed
  • Access to publish/run MCP servers (Claude Code configuration or your MCP client)

Installation steps:

  1. Install the MCP server via npx in your environment (no global install required):
# Bot token (recommended for Slack apps)
npx -y @pilat/mcp-notify

or configure via Claude Code using the example config below (Bot token):

{
  "mcpServers": {
    "mcp-notify": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@pilat/mcp-notify"],
      "env": {
        "SLACK_MCP_BOT_TOKEN": "xoxb-..."
      }
    }
  }
}
  1. If you prefer the user-token mode, supply the user token and cookie token in the environment:
SLACK_MCP_XOXC_TOKEN=xoxc-...
SLACK_MCP_XOXD_TOKEN=xoxd-...
npx -y @pilat/mcp-notify
  1. If using the MCP via a YAML/JSON config (Claude Code), ensure the configuration is saved to your MCP config file (e.g., ~/.claude.json or .mcp.json):
{
  "mcpServers": {
    "mcp-notify": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@pilat/mcp-notify"],
      "env": {
        "SLACK_MCP_BOT_TOKEN": "xoxb-...",
        "SLACK_MCP_XOXC_TOKEN": "xoxc-...",
        "SLACK_MCP_XOXD_TOKEN": "xoxd-..."
      }
    }
  }
}

Additional notes

Tips and caveats:

  • Do not set both bot token and user token environment variables simultaneously; the server expects one mode of authentication at a time.
  • The data directory can be customized with SLACK_MCP_DATA_DIR; ensure the path is writable by the MCP process.
  • The server uses a lightweight SQLite cache (WAL mode) to improve performance and enable concurrent access; monitor disk usage if you run many instances.
  • For user-token mode, ensure you secure the xoxc token and the cookie value (d= token) as described in the README; exposure could allow misuse.
  • If you need to update configuration, restart the MCP server to pick up new environment variables or token changes.
  • This MCP focuses on a single send_message tool; if you need advanced features like reading channels or reacting, consider the full-featured Slack MCP server linked in the README.

Related MCP Servers

Sponsor this space

Reach thousands of developers