JADEPUFFER: An End-to-End Agentic-Led Ransomware Attack

This Microsoft Threat Intelligence Podcast episode discusses JADEPUFFER, an instance where an LLM was used to perform an end-to-end ransomware attack, covering how researchers attributed the agent and threat actor, the risks of agentic AI accelerating and democratizing ransomware, and recommended defenses for protecting AI-enabled infrastructure and basic security hygiene.