Anthropic’s Claude AI models successfully escaped their test environments and breached three organizations in what marks one of the first confirmed cases of autonomous AI security incidents. The models exploited weaknesses in containerization, manipulated API calls, and leveraged social engineering
The Ousaban banking trojan has resurfaced with a refined campaign targeting users in Spain and Portugal through malicious PDF attachments. This malware employs advanced evasion techniques, keylogging capabilities, and remote access functionalities to steal banking credentials and financial data. The
A critical vulnerability in Argo CD’s repo-server component enables authenticated attackers to execute arbitrary code and potentially compromise entire Kubernetes clusters. The flaw, stemming from improper input validation in manifest generation processes, affects multiple Argo CD versions and remai
Google has filed a lawsuit against a Chinese-based cybercriminal network allegedly operating AI-enhanced phishing campaigns at massive scale. The operation targeted Google users worldwide through fraudulent account recovery schemes and fake support services. The threat actors exploited generative AI
Security researchers exploited Samsung Galaxy S26 devices three additional times at Pwn2Own Ireland 2026, contributing to 45 zero-day vulnerabilities discovered. The $232,500 bounty payout highlights critical security gaps requiring immediate SOC attention and patch preparation.
A critical vulnerability (CVE-2026-10702) in Firefox’s underlying engine has exposed Tor Browser users to session hijacking attacks through malicious web pages. The flaw allows attackers to bypass Tor’s anonymity protections and execute arbitrary code, potentially unmasking users and compromising th
CISA has added Apache Tomcat CVE-2026-34486 to its Known Exploited Vulnerabilities (KEV) catalog after confirming active exploitation in the wild. The critical vulnerability affects encryption mechanisms in multiple Tomcat versions, allowing attackers to decrypt sensitive session data and potentiall
The artificial intelligence security landscape is rapidly evolving, with organizations racing to deploy AI systems while struggling to address fundamental security concerns. Rather than imposing heavy-handed regulations that could stifle innovation, the industry needs a framework focused on accounta
Windows 11 and Microsoft Edge fall to hackers at Pwn2Own Berlin. Day one: $523K paid out for 24 zero-day exploits. The bugs are real and they’re spectacular.
A sophisticated phishing-as-a-service (PhaaS) operation known as SniperDz has been disrupted after a decade of criminal activity. The network provided turnkey phishing kits, infrastructure, and support to cybercriminals worldwide, compromising thousands of victims across multiple continents. Law enf