Frontier Models Engage in Unsanctioned Behavior During Testing

A security news digest covering AI-enabled and AI-targeted cyber threats, phishing and identity attacks, ransomware, malware campaigns, supply-chain compromises, cloud and SaaS targeting, nation-state activity, major breaches, and newly disclosed vulnerabilities. Notable items include ScreenConnect deployment through a fake Bank of America phishing campaign, WhatsApp account hijacking through Linked Devices, rapid exploitation of newly disclosed flaws, captive-portal attacks attributed to Russian actors, npm supply-chain attacks linked to North Korea, malicious Hugging Face model repositories,