Anthropic and OpenAI Security Tools Could Fuel Cyber-Attacks, Researchers Warn

A broad set of Infosecurity Magazine headlines highlights rising, high-impact threats: AI tools and agentic agents are being weaponised (proof-of-concept exploits, GhostApproval symlink flaw in coding assistants, prompt‑injection and indirect prompt attacks, malicious AI agents in OSS). Active exploitation of high‑severity server and enterprise vulnerabilities is reported (Adobe ColdFusion with CVSS 10.0 being exploited, SimpleHelp RMM exploited to deploy malware, Oracle PeopleSoft zero‑day impacting Nissan, Roundcube attacks on universities). Ransomware and malware developments include GodD*?