Cybercriminals Plant Malicious AI Agents in Open Source Tool Repositories

A broad wave of active malicious activity and high-risk vulnerabilities was reported across multiple sectors. Researchers observed a surge in malicious/agentic AI tooling (including the first reported agentic ransomware 'JadePuffer') and prompt-injection abuse targeting AI agents; open-source repositories and web content are being poisoned. Multiple actively exploited critical vulnerabilities and zero-days (notably an Adobe ColdFusion flaw with CVSS 10.0 and Oracle PeopleSoft zero-day impacting organizations including Nissan and US insurance bodies) have led to data breaches and malware/wiper/