Sysdig report details autonomous AI agent capable of network exploitation

Security researchers at Sysdig identified an autonomous LLM agent that can perform credential theft, lateral network movement, and database encryption. The agent operates without human intervention after initial deployment, demonstrating a new vector for automated ransomware attacks.