Anthropic Says Claude Mistook the Open Internet for a CTF and Breached Three OrganizationsThe Hacker News·Jul 31, 16:17 UTC · Jul 31, 2026Data breach160
Anthropic Finds Claude Breached Real Companies During Security EvaluationsSecurity Affairs·Jul 31, 08:55 UTC · Jul 31, 2026Data breach45
Anthropic's Claude breached three companies during security testsHelp Net Security·Jul 31, 00:00 UTC · Jul 31, 2026Data breach45
⚡ Weekly Recap: Rogue AI Models, $88M Bitcoin Theft, WaterThe Hacker News·Aug 3, 14:03 UTC · Aug 3, 2026Data breachCVE-2026-42897CVE-2026-6606660
Claude published malicious code to the Internet and attacked 3 real companiesArs Technica · Security·Jul 31, 20:39 UTC · Jul 31, 2026Data breach160
AI agent deception moves from theory to reality in UK cyber testsHelp Net Security·Aug 5, 00:00 UTC · Aug 5, 2026Data breach57
Irregular, firm behind AI hacking incidents, won't say if there were moreThe Record·Aug 7, 15:06 UTC · Aug 7, 2026Data breach45
Week in review: Acrobat Reader flaw exploited, Claude Mythos offensive capabilities and limitsHelp Net Security·Apr 19, 00:00 UTC · Apr 19, 2026Data breach in the wildCVE-2026-34621CVE-2026-39813CVE-2026-3980860