What is it all about?
Overview:
Guardz offers a native MCP (Model Context Protocol) connector for Claude. Once connected, you can query your Guardz data, generate reports, and ask for recommendations — all from within their own Claude workspace, using live Guardz data as context.
The connector is read-only: it lets admins retrieve and summarize information from Guardz, but it cannot make changes to your account or environment.
This feature is available on any plan.
What Claude Can Access:
Through the Guardz MCP, your AI agents can see the customers you have access to in Guardz. That covers each customer's posture score, open issues, security incidents, devices running the Guardz agent, and users. It can also compare customers side by side and link back to each one in the portal.
Once you're connected, you can use Claude's connector settings to choose which Guardz data Claude can access. You can also decide whether Claude uses it automatically or asks for your approval first.
Setting up the Connector
Log in to Guardz platform.
Log in to Claude as admin (Team & Enterprise Plans).
In Claude, go to Settings > Connectors > Add custom connector.
Name it (e.g. Guardz MCP) and enter one of the following Guardz Server URLs (according to your region):
Leave authentication set to Always Required, and under OAuth Client select the following Client ID: DmDilraIo3Urlp00nc3aeVuWcYIM6bgH
Client Secret: leave blank.
Click Add.
When prompted, log in with your Guardz account credentials to complete the connection.
No setup is required on the Guardz platform side — the entire process happens in Claude's connector settings, and you're only prompted to log in to Guardz once, to authorize the connection.