By Vero SecurityBuilt and funded by Vero Security, the AI Security Engineering platform.Read the charter

Treating MCP tool annotations as untrusted hints

MCP tools can carry annotations that describe behavior. The specification says clients must consider them untrusted unless they come from trusted servers. Do not auto approve a tool because it claims to be read only. Show tool inputs to the user, validate results, and log usage.

Read in this order

  1. OfficialTools (Model Context Protocol specification, 2026-07-28)The current spec text: clients must treat tool annotations as untrusted unless the server is trusted.

Save your progress

We email you a sign-in link. No password.

By signing in you agree to the terms. The privacy policy explains how we use your details.

Join the AI security engineering community

Free. Roadmap updates, chapters near you, and first word when credentials open.

We’ll email you community updates. Unsubscribe anytime. By sending this you agree to the terms. The privacy policy explains how we use your details.