The now-patched vulnerabilities in the rapidly growing AI agent framework allow attackers to steal credentials, escalate privileges, and maintain persistence.

This is an ingested summary. Full reporting is available at the primary source below.