What it does
VantaGate intercepts actions an AI agent flags as high-risk, sends them to a human approver through Slack or Email, and returns a signed decision so the agent can resume or stop. The MCP server exposes two tools, create_vantagate_checkpoint and check_vantagate_status, and the publisher also describes an OpenAPI integration. Decisions are delivered to a callback_url as signed HTTPS POST requests, retried up to five times with exponential back-off, and signatures use HMAC-SHA256.
It suits workflows where an agent could send messages to many users, delete records, move money or deploy to production, and a person should approve first. The README lists setup for Claude Desktop, Cursor, Cline and Windsurf, and a local build for development.
Setup requires a VantaGate account and a project API key, which is added to the MCP client configuration. Webhook signatures are checked with a separate signing secret from the project settings, and the README includes a Node.js verification example. Per the publisher, the API key is hashed server-side, payload data is encrypted with AES-256 and destroyed after the decision, and callback URLs are checked against private IP ranges.
Limitations and requirements
Requires a VantaGate account and an API key from the dashboard. Webhook verification uses a separate signing secret found in Project Settings. Slack delivery requires a Slack webhook URL supplied per request. Timeout values are capped at the maximum for the account's plan, for example 24h on the Free plan. The publisher states that signatures should always be verified.
Safety
Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.
- Runs code or changes an assistant's behaviour
Sources
- Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
io.github.Aderix/vantagate-mcp-server
VantaGate MCP Server & OpenAPI integration - Human-in-the-Loop for AI Agents
v1.0.2As shown on registry.modelcontextprotocol.io when we last read it