SLACKMCP Server
Team messaging and notifications via your agent. Read channels, post messages, manage threads, and automate team communication.
WHAT IT DOES
The Slack MCP Server lets your AI agent read and write messages in your Slack workspace. You can pull context from channel conversations, post status updates, reply in threads, and search message history — all without leaving your coding environment. It is especially useful for workflows where you need to cross-reference a Slack discussion with your code, or when you want to notify your team about a deployment or completed task directly from your agent.
To set up, create a Slack app at api.slack.com/apps and configure it with the Bot Token scopes you need (at minimum: channels:read, channels:history, chat:write, and users:read). Install the app to your workspace and copy the Bot User OAuth Token. Add the server to your MCP config and set the SLACK_BOT_TOKEN environment variable. Run npx @anthropic/mcp-server-slack to test the connection. Note that Cursor and Windsurf support is limited, so verify functionality if you are on those clients.
Practical prompts makers use daily: "Read the last 20 messages in #engineering-incidents and summarize what happened with the database outage." Or "Post to #deployments: 'v2.4.1 shipped to production — includes the cart fix and rate limiter update.'" Thread management is where it gets powerful: "Find the thread in #backend where someone asked about the Redis caching strategy and reply with the approach we decided on in our code review." The agent handles the Slack API complexity so you can focus on the actual communication.