VulnerabilitiesCSS Attacks in Webmail Can Steal Tokens and Manipulate AI Agents
Research published on August 9, 2026 by Gareth Heyes, a PortSwigger researcher, describes attacks that can be carried out using only HTML/CSS against
Vulnerabilities — Page 15
More news
Page 15 / 21 · 188 articlesVulnerabilities
VulnerabilitiesChina Launches Cybersecurity Review of Palo Alto Networks Products
China initiates cybersecurity review of Palo Alto Networks products to protect critical infrastructure, amid geopolitical tensions and tech self-reliance push.
VulnerabilitiesWindows Hello for Business Can Be Abused by Malware to Maintain Access to Entra ID
Malware can exploit Windows Hello for Business to maintain persistent access to Microsoft Entra ID without admin privileges. Learn about detection and mitigations.
VulnerabilitiesHead Mare Compromises TrueConf Servers and Distributes Clients with PhantomCore
Head Mare compromises TrueConf servers, distributing PhantomCore malware via malicious clients. Vulnerable versions need immediate updates.
VulnerabilitiesCSS Attacks in Webmail Can Steal Passwords and Tokens—and Direct AI Agents
Discover how CSS attacks in webmail can compromise passwords, tokens, and AI agents, as revealed at Black Hat USA 2026, affecting major email services.
VulnerabilitiesHTTP Terminator Uses AI to Discover New HTTP Desynchronization Techniques
PortSwigger has introduced HTTP Terminator , an AI-assisted research system developed by James Kettle to generate and validate HTTP desynchronization
VulnerabilitiesCritical Vulnerability in Kemp LoadMaster Appliances: 792 Exploitation Attempts Detected
Critical Kemp LoadMaster flaw CVE-2026-8037: 792 exploitation attempts detected. Unauthenticated command injection; apply patches urgently.
VulnerabilitiesMetabase Zero-Day SQL Injection Exploited to Steal Data from Customer Instances
Critical zero-day SQL injection in Metabase allows data theft from cloud and self-hosted instances. Updates and mitigation steps provided.
VulnerabilitiesVulnerabilities in CPDLC Communications: Fake Messages and Radio-Induced Disconnections
Learn about CPDLC vulnerabilities in ATN-B1 systems, including fake messages and disconnections from CISA advisory ICSA-26-219-01. CVE details included.