untrustedContentHint
A safety annotation in the WebMCP protocol that marks user-generated content or external data returned by a tool as untrusted, helping the agent identify and defend against indirect prompt injection attacks.
Timeline (last 90 days)
The Chrome team demonstrated three security annotations (hints) for WebMCP tool definitions: readOnlyHint, consequentialHint, and untrustedContentHint.
The untrustedContentHint: true annotation marks user-generated content or external data returned by a tool as untrusted, to defend against indirect prompt injection.
All Facts (2)
The Chrome team demonstrated three security annotations (hints) for WebMCP tool definitions: readOnlyHint, consequentialHint, and untrustedContentHint.
50%UnverifiedThe untrustedContentHint: true annotation marks user-generated content or external data returned by a tool as untrusted, to defend against indirect prompt injection.
50%