Slack MCP Server — 让 AI Agent 读取、发送、整理 Slack 消息
Slack MCP 服务器让你的 AI 编程助手变成 Slack 原生成员。在 Claude Code、Cursor 或任何 MCP 兼容客户端中读取频道历史、发布更新、回复线程、跨工作区搜索。
安装 Slack MCP Server
{
"mcpServers": {
"slack": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-slack"],
"env": {
"SLACK_BOT_TOKEN": "xoxb-your-bot-token",
"SLACK_TEAM_ID": "T01234567"
}
}
}
}Slack MCP Server — Slack Integration for AI
MCP server for Slack workspace integration. Read channels, post messages, search conversations, and manage threads from Claude Code or Cursor.
Slack MCP — Team Messaging Server for AI Agents
MCP server that connects AI agents to Slack workspaces. Send messages, read channels, search history, manage threads, and automate team communication via natural language. 1,500+ stars.
Airweave — Context Retrieval Layer for AI Agents, 50+ Integrations
Open-source context retrieval layer connecting AI agents to 50+ apps including Notion, Slack, GitHub, and Jira. Unified search API with MCP support.
Composio — 250+ Tool Integrations for AI Agents
Composio connects AI agents to 250+ tools (GitHub, Slack, Jira, DBs) with managed auth. 15K+ stars. Python/JS SDK, MCP support. AGPL-3.0.
为什么每个 AI Agent 都需要 Slack
The Slack MCP server closes the oldest gap in AI-assisted development: your AI writes code, your team discusses the code in Slack, and the AI never sees the conversation. Install this one integration and your agent can read the #engineering channel, pull up last week's incident retro, respond to a direct message with a build-status summary, or post a PR review into the right thread automatically.
What makes Slack-over-MCP powerful is bidirectional context. A code-review agent no longer needs to be told "look at PR #123 — the customer reported flakiness in the login flow last Tuesday." It can search Slack for "login flakiness" and pull the original report itself. A release agent can post a changelog into #releases, tag the on-call engineer, and react to the thread when CI finishes. For autonomous agents, this is the difference between a disconnected tool and a team member.
The server respects Slack's permissions model: it only sees channels the bot user has been added to, and can only post as the bot. Combine it with a database MCP server and you get agents that answer Slack questions with live data. Pair with the GitHub MCP server and your team gets automated, in-Slack PR reviews with zero custom glue code.
Slack is where your team thinks out loud. An AI agent that can't read Slack is working with half the context.
常见问题
What is the Slack MCP server?+
The Slack MCP server is an open-source MCP implementation that wraps the Slack Web API — channels, messages, search, users, reactions — into tools an AI agent can call. It runs locally alongside your MCP client (Cursor, Claude Code, Windsurf, etc.) and authenticates to Slack using a standard bot token.
How do I create a Slack bot token for the MCP server?+
Go to api.slack.com/apps, create a new app, pick 'From scratch', choose your workspace, then under 'OAuth & Permissions' add the scopes: channels:history, channels:read, chat:write, search:read, users:read. Install the app to your workspace and copy the 'Bot User OAuth Token' (starts with xoxb-). That's your SLACK_BOT_TOKEN. Your Team ID is visible in the URL when you open your workspace in a browser (the Txxxxx part).
Can the Slack MCP server post to private channels?+
Yes, but only to channels the bot has been invited to. Slack's permission model means even with channels:read scope, private channels are invisible until a human user runs /invite @your-bot in the channel. This is a feature, not a bug — it keeps AI access scoped to what the team has explicitly approved.
Does it work with Slack Enterprise Grid?+
Yes, but each workspace in the grid needs its own bot installation, or you can build an Enterprise-level Slack app with admin.* scopes for cross-workspace access. For most teams, single-workspace bot tokens are simpler and adequate.
What can I build with the Slack MCP server?+
Common patterns: (1) An on-call agent that pulls recent alerts from #alerts, correlates them with git deploys, and posts a root-cause hypothesis. (2) A release agent that generates changelogs from PRs and posts them into #releases. (3) A support triage agent that reads #customer-support, categorizes issues, and files tickets. (4) A meeting notes agent that reads DM threads and summarizes them into your personal notes doc. Browse TokRepo's <a href="/en/ai-tools-for/agents">agent-building tools</a> for more templates.