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

The feed reports active exploitation of multiple critical vulnerabilities, malware campaigns, supply-chain compromises, credential theft, botnet activity, and AI-assisted attack techniques. Notable incidents include zero-day exploitation of SonicWall SMA 1000, unauthenticated RCE in Switchvox and GeoNetwork, rapid exploitation of a JFrog Artifactory authentication bypass, a BGP-hijacked Virtualizor update causing root compromise, and malicious Git configurations that trigger code execution by AI coding agents. Additional reporting covers Android banking malware, RATs, ransomware, fake software