Experts Warn of JADEPUFFER: First Documented LLM-Driven Agentic Ransomware Attack

Summary

Cybersecurity experts are warning about JADEPUFFER, identified as the first documented case of agentic ransomware, which is entirely run by a Large Language Model (LLM). Researchers at Sysdig discovered this dangerous attack exploited a known vulnerability, adapted to obstacles, and targeted an Alibaba Nacos instance, notably failing to back up data even if a ransom were paid.