Find articles
24 results for “agentic”
Storm-3358 Deploys ARTEX AI Framework Against Financial Sector
Microsoft Threat Intelligence observed financially motivated actor Storm-3358 using the agentic AI offensive framework ARTEX to target financial organizations in October 2026.
Google Cloud Security Summit: Agentic-Era Defenses from CodeMender to Auth Manager
Google Cloud unveiled machine-speed security tools at its Security Summit, including CodeMenter for autonomous vulnerability remediation, Auth Manager for agent identity, and expanded VPC Service Controls for identity-based perimeters.
Amazon WorkSpaces for AI Agents Adds Linux and Windows 11 Desktops
WorkSpaces for AI agents now runs on Linux and Windows 11 alongside Windows Server, letting agents operate in the same OS environments as developers and end users without re-platforming workloads.
Patch Node.js 22, 24, 26 in Azure DevOps pipelines after October 2026 security release
Node.js 22.23.4, 24.22.1, and 26.12.1 fix DNS, TLS, and Brotli flaws. Update build agents and runtime tasks now.
LLMJacking Surges as Threat Actors Monetize Stolen Cloud Model Access
Google Threat Intelligence Group details how financially motivated clusters hijack enterprise LLM APIs, generating six-figure daily bills and reselling access through underground LLM Mixers.
Gateway API v1.6 Graduates TCPRoute and UDPRoute to Standard
Gateway API v1.6 promotes TCPRoute and UDPRoute to v1 stability for layer 4 routing, while moving experimental resources like XBackend into a separate API group with X-prefixed names.
GKE API timeouts traced to three bugs across PSC, Envoy, and Linux kernel
A multi-week outage blocking 800 developers across europe-west clusters was caused by packet reordering in Private Service Connect, Envoy half-close behavior, and a Linux kernel bug that silently dropped FIN packets in FIN_WAIT_2 state.
Scale GKE Networking for High-Density AI Agent Workloads with Dataplane V2
GKE Dataplane V2 now handles millions of concurrent AI agent sandboxes by separating identity labels from metadata annotations and filtering high-cardinality labels, enabling sub-second provisioning at 300 sandboxes per second.
The CISO's New Mandate: Governing Autonomous Agents Without Slowing Innovation
As AI agents gain autonomy, CISOs must shift from perimeter defense to identity-centric governance and AI-powered threat detection to secure the agentic enterprise.