Gartner says AI SOC agents are moving rapidly from hype into real-world evaluations, but most organizations will only see measurable gains if they use a structured, outcomes-driven framework. The report highlights seven key evaluation areas, including workload reduction, vendor durability, analyst impact, autonomy boundaries, stack integration, and transparency, while survey data shows broad adoption but cautious trust. #Gartner #AI_SOC_Agents #Prophet_Security

Tencent’s Zhuque Lab released AI-Infra-Guard, an open-source scanner for AI systems that fingerprints services like Ollama, vLLM, and ComfyUI, checks thousands of CVEs, inspects MCP servers and agent skills, and tests models for jailbreak resistance. It also highlights risks from indirect prompt injection and notes that the tool should not be exposed directly because the open-source build has no built-in authentication. #AI-Infra-Guard #ZhuqueLab #Ollama #vLLM #ComfyUI #SkillTrustBench #MCP

CIA cyber operators helped build a “flawless intelligence picture” that enabled U.S. forces to locate and capture Nicolás Maduro and his wife in Caracas during Operation Absolute Resolve. Deputy Director Michael Ellis also said AI will increasingly shape CIA cyber operations, intelligence collection, and analysis, with human oversight and safeguards in…

Malone Lam, a Singaporean national known as “Anne Hathaway” and “$$$,” pleaded guilty to racketeering for leading the Social Engineering Enterprise, which stole more than $245 million in cryptocurrency through coordinated scams and home break-ins. Prosecutors said the group impersonated Apple, Google, and customer support staff, targeted wealthy crypto holders, and…

CISA has added CVE-2026-86218, a critical static code injection flaw in N-able N-central, to its KEV catalog and ordered FCEB agencies to patch it by September 11, 2026. N-able said the issue has been exploited in the wild, while Huntress is investigating a compromise of a fully patched N-central environment and…

Researchers at Calif discovered and privately reported a critical WeChat vulnerability that enabled them to build “WeWorm,” a worm that spreads through WeChat calls without any user interaction. The flaw let attackers compromise accounts and propagate across iOS and Android devices, prompting Tencent to release mitigations for the app and its servers. #WeChat #Tencent #WeWorm #Calif

Microsoft patched 974 defects in its largest-ever Patch Tuesday update, including two actively exploited zero-day vulnerabilities in the Windows Update Stack and Windows Advanced Local Procedure Call. Security experts said AI-assisted vulnerability discovery is increasing the number of disclosed flaws, but it has not yet caused a matching surge in real-world exploitation. #CVE-2026-81963 #CVE-2026-85880 #WindowsUpdateStack #WindowsAdvancedLocalProcedureCall

Microsoft’s latest Patch Tuesday set a new record with 973 disclosed bugs, including two actively exploited vulnerabilities tracked as CVE-2026-81963 and CVE-2026-85880. Researchers warned that thousands of unpatched Exchange servers and weaknesses in the Windows update stack could help attackers chain exploits into ransomware-style intrusions. #CVE-2026-81963 #CVE-2026-85880 #Microsoft #Windows #Exchange #AdobeCommerce…