Google, Anthropic, and OpenAI Unveil Cyber AI Models, Safeguards, and Access Programs

The document is a cybersecurity news digest covering active exploitation, malware campaigns, supply-chain and infrastructure attacks, AI-agent security flaws, threat actors, and law-enforcement disruption activity. Highest-impact items include zero-day exploitation of SonicWall SMA 1000 VPN appliances, unauthenticated RCE in GeoNetwork and Switchvox, rapid exploitation of an authentication-bypass flaw in JFrog Artifactory, a BGP hijack delivering a malicious Virtualizor update with root access, and malware campaigns targeting Windows, Android, iOS, Linux, and macOS users. The reporting also──