Gong and other scholars have been issuing warnings about the security vulnerabilities of AI agents for a while. They publish ...