Afternoon Review in IT Security — September 11, 2026

Thursday's security landscape reflects an escalating threat environment where critical infrastructure vulnerabilities, AI-powered attack automation, and supply chain compromises demand immediate organizational response. Four significant incidents underscore the convergence of traditional exploitation techniques with emerging AI-driven attack methodologies.

GitLab Urges Users to Patch Maximum-Severity Path Traversal Flaw

GitLab has issued an urgent call for immediate patching against a maximum-severity path traversal vulnerability tracked as CVE-2026-85706. The flaw represents a critical risk to GitLab server deployments and requires administrators to prioritize remediation efforts to prevent unauthorized file system access and potential system compromise. Source: GitLab urges users to patch max severity path traversal flaw

Anthropic Reveals Russian Hackers Leveraging Claude AI for Malware Evasion

Anthropic has disclosed that Russian criminal groups are actively targeting AI vendors' infrastructure, including attempts to steal pre-release versions of the Claude model. The attackers have weaponized Claude AI itself to automate malware evasion techniques, demonstrating a sophisticated approach where adversaries exploit AI capabilities to enhance their operational effectiveness. This revelation highlights the dual-use nature of advanced AI systems and the emerging threat model where threat actors target AI vendors as both infrastructure targets and toolsets for attack automation. Source: Anthropic Says Russian Hackers Used Claude AI to Automate Malware Evasion

Russian Threat Actor Employs AI to Exploit PaperCut Vulnerabilities at Scale

A Russian threat actor has leveraged artificial intelligence to build, test, and deploy exploits against PaperCut print management systems across hundreds of organizations worldwide. The campaign demonstrates how adversaries are integrating AI-powered development and testing capabilities into their exploitation workflows, enabling rapid iteration and deployment at unprecedented scale. This convergence of AI automation with traditional vulnerability exploitation represents a significant evolution in threat actor capabilities. Source: PaperCut Flaws Exploited in AI-Powered Attacks

Attackers Chain JFrog Artifactory Vulnerabilities to Achieve Administrative Control

Attackers have successfully chained two vulnerabilities in JFrog Artifactory to gain administrator control of self-hosted servers and establish persistent backdoors, according to research from cloud security firm Wiz. The attacks were observed between August 15 and September 8, targeting organizations that had failed to apply available patches. Since JFrog had released fixes for both flaws prior to the observed campaign window, only unpatched deployments remained vulnerable to this chained exploitation technique, underscoring the critical importance of timely patch deployment in software repository infrastructure. Source: Attackers Chain JFrog Artifactory Flaws to Gain Admin Control and Plant Backdoors

Today's threat intelligence indicates that organizations must prioritize patch management across development infrastructure, monitor for AI-augmented attack campaigns, and implement defense-in-depth strategies for critical software supply chain components. The integration of AI capabilities into threat actor workflows represents a qualitative shift in the threat landscape requiring elevated vigilance and accelerated security response protocols.