AI Agents, Open-Weight Models, and LLM-Targeted Ads: Three Emerging Risks for Security Leaders
A Black Hat post-mortem on AI agent behaviour, China's accelerating open-weight parity, and the first…
Month: August 2026
CSS-Based Attacks Found to Break Webmail Boundaries Across Major Providers
Research from PortSwigger demonstrates that malicious CSS inside an email can escape its rendering sandbox…
N-able confirms customer network breaches as second mandatory hotfix lands for N-central zero-day
Attackers exploited a critical authentication bypass in N-central to pivot from the management platform into…
OpenAI’s Training Agents Accidentally Attacked Hugging Face — A Full Timeline
A Black Hat presentation reveals how autonomous AI agents, given no internet access, improvised their…
WordPress Login Screen XSS Vulnerability Opens Path to Server-Side Code Execution
A pre-authentication flaw in WordPress affects all versions of the platform and can be chained…
Thousands of Industrial Water Controllers Remain Exposed Online After Wave of US Utility Attacks
A Forescout scan found over 4,000 Rockwell Automation and Allen-Bradley controllers accessible on the public…
Google Patches Agent-to-Agent Trust Boundary Flaws in Python APK
Vulnerabilities in a Google Python package allowed a lower-privilege AI agent to exploit a higher-privilege…
CISA Orders Federal Agencies to Patch Langflow, N-central and Apache Tomcat Within Three Days
Three actively exploited vulnerabilities across widely deployed platforms are now on CISA's Known Exploited Vulnerabilities…
AI Election Tools Are Improving but Remain Unreliable for Voters
New research finds error rates dropping to zero in some tests, yet critical gaps in…
CISA Adds Three Actively Exploited Vulnerabilities to KEV Catalog
IBM Langflow, N-able N-central, and Apache Tomcat vulnerabilities are confirmed in active use by threat…