Slack Enterprise MCP Server

Enterprise-grade Slack integration with compliance audit trails for AI agents. Every action is logged locally for governance, regulatory compliance, and security review.

Built by MEOK AI Labs -- the team behind MEOK AI OS (22 APIs, 15 AI models, 307 tests).

Why this exists

Existing Slack MCP servers send messages and search -- but none provide the audit trail that enterprise teams require. This server logs every action with timestamps, caller identity, and operation details to a local JSONL file and in-memory buffer, enabling compliance review without external dependencies.

Tools

Tool	Description
`send_message`	Send message to channel or thread (audit-logged)
`search_messages`	Search across workspace with Slack modifiers
`list_channels`	List channels with member counts and topics
`get_thread`	Get full thread with all replies
`create_channel`	Create new channel with description
`set_channel_topic`	Update channel topic
`get_audit_log`	Retrieve audit trail of all MCP actions (enterprise feature)
`summarize_channel`	Get last N messages with participant and activity breakdown

Installation

pip install mcp slack_sdk

Configuration

Set the SLACK_BOT_TOKEN environment variable. Create a Slack app at https://api.slack.com/apps with the following Bot Token Scopes:

chat:write -- send messages
channels:read -- list channels
channels:history -- read channel history
groups:read -- list private channels
groups:history -- read private channel history
search:read -- search messages

Optional environment variables:

Variable	Default	Description
`SLACK_BOT_TOKEN`	(required)	Slack Bot User OAuth Token
`AUDIT_LOG_PATH`	`audit_log.jsonl`	Path for the audit log file

Usage

Run the server

SLACK_BOT_TOKEN=xoxb-your-token python server.py

Claude Desktop config

Add to ~/.claude/claude_desktop_config.json:

{
  "mcpServers": {
    "slack-enterprise": {
      "command": "python",
      "args": ["/path/to/slack-enterprise-mcp/server.py"],
      "env": {
        "SLACK_BOT_TOKEN": "xoxb-your-token"
      }
    }
  }
}

Example calls

Send a message:

Tool: send_message
Input: {"channel": "#general", "text": "Deployment complete for v2.3.1"}
Output: {"ok": true, "channel": "C01ABC123", "ts": "1234567890.123456"}

Search workspace:

Tool: search_messages
Input: {"query": "deployment error in:#engineering from:@alice", "count": 10}
Output: {"total": 3, "messages": [...]}

Get audit trail:

Tool: get_audit_log
Input: {"limit": 20, "action_filter": "send_message"}
Output: {"entries": [{"timestamp": "2026-04-13T...", "action": "send_message", ...}], "total_logged": 142}

Audit Log Format

Each line in the JSONL audit file contains:

{
  "timestamp": "2026-04-13T10:30:00.000Z",
  "action": "send_message",
  "caller": "anonymous",
  "params": {"channel": "#general", "text_length": 42},
  "result_summary": "Sent to #general, ts=1234567890.123456"
}

Sensitive data (tokens, full message bodies) is excluded from audit logs by design.

Security Considerations

Token storage: Never commit SLACK_BOT_TOKEN to version control. Use environment variables or a secrets manager.
Audit log access: The audit log file contains metadata about all operations. Restrict file system access to authorized users.
Rate limiting: Built-in rate limiting prevents abuse. Free tier allows 100 calls/day; Pro tier allows 10,000 calls/day.
Data minimization: Message text is truncated in audit logs. Full content is not persisted in the audit trail.
Principle of least privilege: Request only the Slack scopes your use case requires.

Pricing

Tier	Limit	Price
Free	100 calls/day	$0
Pro	10,000 calls/day + priority	$12/mo
Enterprise	Custom + SLA + on-prem audit	Contact us

License

MIT

Slack Enterprise MCP Server