HiddenLayer Automated Red Teaming prevents malicious manipulation of AI modelsHelp Net Security·Nov 21, 00:00 UTC · Nov 21, 2024AI safety & security30
Cursor AI Code Editor Fixed Flaw Allowing Attackers to Run Commands via Prompt InjectionThe Hacker News·Aug 2, 05:40 UTC · Aug 2, 2025AI safety & securityCVE-2025-54135135
New TokenBreak Attack Bypasses AI Moderation with SingleThe Hacker News·Jun 12, 15:41 UTC · Jun 12, 2025Model release150