Security leads today’s list. Google’s Threat Intelligence Group documents the first fully autonomous AI-agent exploit chain observed in the wild, Amazon details ARTEX campaigns that steal AI credentials at scale, and Canonical begins offering AI-generated security patches to Ubuntu Pro. From the labs, OpenAI announces its first phase of embedded independent safety assessments. Beyond AI: an 11-year-old Linux kernel TCP flaw that allows unauthenticated remote crashes, and Apple’s tightened macOS permission controls.
From the labs
Worth your time
- GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold — Google Cloud (first fully autonomous AI-agent exploit chain) (+ The Register, The Hacker News)
- Amazon Threat Intelligence observes threat actors running cybercrime campaigns with ARTEX AI agents — AWS Security Blog (+ The Register, FBI Flash alert)
- New Edge archive offers opt-in AI-driven security remediation and backports — Canonical (Surge agents patch Ubuntu Pro) (+ The Register)
Beyond AI
- NoisyTCP: Linux Kernel TCP Pre-auth Remote Corruption (CVE-2026-99288) — Hyunwoo Kim (+ The Register, PoC and advisory repo)
- Updates to Device Control and Data Access, and Full Disk Access in macOS — Apple (macOS 27.3 beta 2) (+ TechCrunch)