“HalluSquatting: LLM Hallucinations Enable AI Agent Botnets”
A newly disclosed attack technique dubbed “HalluSquatting”...
GitHub Copilot Agents Vulnerable to Workflow-Level Jailbreaks
GitHub Copilot’s new IDE-integrated coding agents are...
Google Chrome 150 fixes 27 vulnerabilities – Update Now
Google has released an urgent security update for Chrome’s...
Foxit Patches Critical PDF Flaws Enabling RCE, Privilege Escalation on Windows
Foxit has issued emergency security updates to remediate a...
IonStack: One-Click Android 17 Root Exploit via Firefox and Linux Flaw
Nebula Security has disclosed a sophisticated exploit chain...
Attackers Clone “Verified” GitHub Commits via Signature Malleability
Git’s long‑standing trust model has been upended by a...
CrowdStrike Finds 5 New Prompt Injection Techniques Targeting AI Agents in 2026
CrowdStrike has identified five new prompt injection...
Claude Cowork Expands to Web, Mobile With Offline AI Task Execution
Anthropic has expanded Claude Cowork beyond its...
16-Year-Old KVM Flaw “Januscape” Exposes Intel, AMD Cloud Hosts to Guest Attacks
A newly disclosed vulnerability, dubbed Januscape...
BeyondTrust Fixes Critical RS, PRA 0-Day Flaws Allowing Auth Bypass, RCE
BeyondTrust has disclosed multiple critical and...