Category: AI Security
Android Expands AI Scam Detection to Samsung Devices and 20+ Countries
Google is rolling out Gemini-powered call and message scam detection to Samsung Galaxy S26 hardware and expanding messaging protections to more than…
How Google Continuously Hardens Workspace Against Indirect Prompt Injection
Google's GenAI Security Team details the layered, iterative pipeline it uses to discover, catalog, and defend against indirect prompt injection attacks targeting…
Google Maps Real-World Prompt Injection Abuse on the Public Web
Google's threat intelligence teams scanned billions of web pages for indirect prompt injection patterns and found a spectrum of abuse ranging from…
Meta’s AI Support Bot Exploited to Reset Instagram Passwords
Pro-Iranian hackers circulated a Telegram tutorial showing how to trick Meta's AI customer support assistant into linking a new email address to…
AI Agent Skills Need Supply-Chain Audits, Unit 42 Research Finds
A new audit primitive called Behavioral Integrity Verification scanned nearly 50,000 agent skills and found that 80 percent deviate from their declared…
AI-Driven Vuln Discovery Has Outpaced Human Patching, Talos Warns
Cisco Talos researcher Yuri Kramarz argues that frontier AI models can autonomously find and exploit zero-days in minutes, collapsing the traditional vulnerability…
Malicious Skills Persist on ClawHub Despite AI Agent Marketplace Scanning
Unit 42 researchers found five undetected malicious skills in OpenClaw's ClawHub marketplace between February and May 2026, including macOS infostealers and novel…
OpenAI Restricts GPT-5.6 Sol Rollout, Touts Stronger Cyber Safeguards
OpenAI has released three GPT-5.6 variants, Sol, Terra, and Luna, in a tightly controlled preview tied to U.S. government engagement, with Sol…
DNS-Delivered Payload Turns Clean GitHub Repos Into AI Agent Traps
Mozilla 0DIN researchers demonstrate how an agentic coding tool can be manipulated into spawning a reverse shell, with no malicious code ever…