Microsoft has warned that agentic AI systems can be weaponized to extract cloud secrets in approximately five minutes, highlighting emerging security vulnerabilities in cloud infrastructure. The warning underscores growing concerns about the potential misuse of autonomous AI agents to compromise sensitive data and credentials stored in cloud environments.