AI & LLM
AI and LLM covers generative models, agents, prompt injection, data security and new artificial intelligence tools. The collection connects capabilities, limitations, operational risks and their impact on technical work.
Langflow Under Attack: MCP Servers Become Vector for Authenticated RCE
CVE-2026-105740 in Langflow has triggered 1,423 exploit attempts on a honeypot. The attack mechanism exploits opaque MCP server config…
From Six Months to Six Days: GPT-5.6 Sol Built the $250,000 Chrome Exploit
XBOW used GPT-5.6 Sol to build a full-chain Chrome exploit in six days with $40,000 in compute. Google paid the maximum $250,000 bount…
Vghost: Tencent's AI Finds Linux Kernel vhost Flaw in Two Hours
CVE-2026-74580, dubbed Vghost, is a Linux kernel vhost subsystem vulnerability discovered with Tencent's Corvus AI multi-agent system.…
Chrome and Firefox: 114 Vulnerabilities Patched as Mozilla Launches Firefox Enterprise
On October 14, 2026, Google and Mozilla released security updates fixing a combined 114 vulnerabilities — 37 in Chrome 155 and 77 in F…
Vercel Sandbox: AI Agents Bypass Network Restrictions via Manipulated Certificates
AIR researchers have demonstrated a vulnerability in the Vercel Sandbox that allows confined AI agents to exfiltrate sensitive data by…
Anthropic: Small AI Models Ready to Self-Replicate on IoT Devices
Anthropic's Frontier Red Team research demonstrates that models up to 30 billion parameters can run distributed inference on $55 TV bo…
AI Completes OT Attack Chains in 16 Minutes: The Booz Allen Test
Two frontier AI models executed eight attack objectives against industrial systems in an isolated lab, moving from perimeter compromis…
Ransomware: Data Exfiltration Surges 275%, Terabytes Become the New Extortion Lever
The 2026 ThreatLabz report documents an explosion in data theft: 896.2 TB exfiltrated, a 275.8% increase. The terabyte replaces encryp…
Cybersecurity and AI Safety: The 24-Hour Playbook the GCRAI Wants to Replicate
A former Intel executive and GCRAI commissioner argues that AI safety must adopt three proven cybersecurity maturities—demonstrable co…
AI Agent Re-grades 10,000 Cloud Alerts: Three-Quarters Had the Wrong Severity
Sola Security study of 9,967 HIGH findings: 58.5% downgraded, 16.6% upgraded. Static CSPM vendor severities do not reflect real risk i…
Phishing Now Strikes Twice: The Human and the AI Assistant
Barracuda Research has documented phishing campaigns that exploit HTML comments, invisible CSS, and zero-width characters to manipulat…
Check Point: Attacks Up 48% YoY as AI Becomes Standard Offensive Tooling
The October 12, 2026 Threat Intelligence Report documents 2,803 weekly attacks per organization and the convergence of generative AI w…