OpenAI’s dirty deeds Down Under included security bypass attempts, using exposed keys, source code siphon

The document is a September 2026 security-news feed covering active exploitation of critical enterprise vulnerabilities, AI-agent-enabled intrusion and malware activity, identity and cloud abuse, phishing and social engineering, ransomware and extortion, cryptocurrency theft, data exposure, and law-enforcement actions. Notable incidents include an actively exploited F5 BIG-IP APM zero-day RCE, critical Citrix/NetScaler vulnerabilities under attack, Salesforce Agentforce zero-click CRM data theft, AI coding-agent RCE, Azure identity hijacking, and campaigns attributed to North Korean and other,