VulnCheck recorded 360 exploitation attempts targeting CVE-2026-0768 in Langflow and Rails KindaRails2Shell CVE-2026-66066 within 72 hours of disclosure.
OpenAI paused internal work on its Astra model after an evaluation found cyber capabilities that may reach a Critical rating
PortSwigger’s AI-assisted HTTP Terminator found roughly 30,000 HTTP desync vectors and a live Apache Traffic Server zero-day affecting roughly 700
Meta admitted its Muse Spark 1.1 model escaped a test environment and breached an unnamed third party’s systems during evaluations
A Claude Mythos 5 agent spent 34 hours trying to merge malware into an open-source project during a UK AI
Google removed three ADK AI workflows after Pillar Security showed a GitHub issue could prompt-inject a triage agent into launching
A Google Firebase misconfiguration in the tl;dv AI meeting tool lets users query others’ meeting data and potentially join calls,
A new postmortem reveals OpenAI’s rogue AI model exploited JFrog Artifactory zero-days to escape its sandbox and breach Hugging Face
Lee Jia Jie used AI assistance to discover CVE-2026-53264, a Linux kernel use-after-free enabling local root escalation. A public exploit
Mindgard researcher Aaron Portnoy disclosed a code execution flaw in Cursor AI editor that silently runs trojanized git.exe files when
Subscribe to the Daily Security Review Email Briefing to stay informed on the latest threats, trends, and technology, along with insightful columns from industry experts.