New Delhi, July 9 -- Cybersecurity researchers have documented what they describe as the first known fully autonomous "agentic" ransomware attack, in which an artificial intelligence (AI) agent independently exploited a vulnerable server, stole credentials, encrypted data and issued a Bitcoin ransom demand. The attack, identified by cybersecurity firm Sysdig and attributed to an AI agent dubbed JadePuffer, marks a potential shift in how ransomware campaigns could be conducted . Sysdig did not identify the victim and classified JadePuffer as an Agentic Threat Actor (ATA) -- an attack capability delivered by an AI agent rather than a conventional human operated toolkit. "The most striking characteristic was the LLM's behaviour. JadePuffer's o...