JadePuffer Returns With Ransomware Designed to Wipe AI Models

Multiple high-risk developments across the threat landscape: a JadePuffer follow-up campaign deployed the ENCFORGE locker designed to destroy AI model artifacts, signalling a new destructive ransomware focus on ML/IP assets. US authorities issued an urgent CISA patch directive for two actively exploited critical Fortinet flaws while Microsoft shipped fixes for a record 570 CVEs, reflecting accelerating discovery and patch volume driven by AI. Other notable threats include a Secure Boot bypass via eleven Microsoft-signed UEFI shims, HollowGraph malware abusing Microsoft 365 calendars/Microsoft