Slack MCP Server: Tools, Scopes, and How to Connect It Safely

It’s Slack’s official remote Model Context Protocol server, hosted at https://mcp.slack.com/mcp. It exposes Slack as tools an AI agent can call — searching messages and files, reading channel and thread history, posting messages, working with canvases, and looking up users. Agents connect over a URL and authenticate with OAuth rather than running anything locally.

Search across channels, group DMs, DMs, files, and users; conversation history for channels, groups, and DMs; message sending; canvas read and write; and user profile lookup. Which are available depends on the OAuth scopes you grant the Slack app, and on TrueFoundry each tool can be individually enabled, disabled, re-described, or annotated.

Yes, with scoping — but the bar is higher than for a code-oriented server, because Slack holds private conversations. Three things do most of the work: per-user OAuth so each agent sees only its user’s channels, trimmed scopes so DM access and posting are never granted unless needed, and human approval on anything that writes.

Yes. TrueFoundry runs in your VPC, on-prem, air-gapped, or hybrid, so prompts and responses never leave your domain even as you route across many providers.

Yes. It includes an MCP Gateway, an Agent Gateway, and an MCP & Agents Registry with tool-level access control. Agents on LangGraph, CrewAI, AutoGen, or a custom framework can all be governed centrally.

Sim. O gateway é compatível com OpenTelemetry e se integra com Grafana, Datadog, Prometheus ou a sua stack preferida. Ele rastreia cada requisição, do prompt à execução da ferramenta e do modelo, para que você obtenha logs unificados sem precisar remover o que você já usa.