Stories cluster related articles into a single narrative, linked to the underlying CVEs and affected products. 3,731 stories synthesized.
Two critical vulnerabilities, DuneSlide (CVE-2026-50548 and CVE-2026-50549), discovered in the AI code editor Cursor, enable prompt injection attacks to bypass security sandboxes and execute arbitrary commands on developer machines.
A new fileless malware campaign dubbed Veil#Drop leverages Google's Blogspot platform to distribute the PureLog Stealer, evading traditional defenses by executing entirely in memory.
A year-old, unpatched vulnerability in Apple's 'Hide My Email' feature allows attackers to reveal users' actual email addresses behind anonymized aliases.
The Ousaban banking trojan, previously targeting Brazil, has been updated to target Spain and Portugal using sophisticated phishing, steganography, and geofencing techniques.
Link11 has unveiled a rebuilt Layer 3/4 DDoS mitigation solution that leverages AI and behavior analysis for rapid threat detection and response.
Critical flaws in the popular Fluentd log collector could allow remote code execution, information exposure, denial-of-service, and server-side request forgery.
SentinelOne's updated Autonomous SOC maturity model emphasizes that achieving AI-driven security operations hinges on robust data foundations and governance, not just advanced tools.
A malicious Google ad campaign is distributing the MacSync Stealer malware via a fake Claude Code CLI installer, aiming to harvest macOS credentials and crypto seed phrases.
A zero-day denial-of-service vulnerability in Anthropic's buffa Rust library, affecting versions prior to 0.8.0, allows attackers to trigger unbounded heap allocations and crash processes.
Threat actors are increasingly leveraging the Bring Your Own Vulnerable Driver (BYOVD) technique to disable antivirus and EDR solutions, gaining kernel-level privileges to bypass defenses.
Citrix has released security updates for its NetScaler ADC and Gateway products, addressing six vulnerabilities, notably a new denial-of-service flaw dubbed 'HTTP/2 Bomb' and a critical information disclosure bug.
Researchers demonstrate how an AI model can transform a theoretical attack concept into a practical browser-only ransomware technique, bypassing traditional security measures.
A widespread campaign is distributing malicious installer archives disguised as popular freeware, leading to the deployment of the ScreenConnect remote access tool and the AsyncRAT malware.
A new phishing operation dubbed ARToken is leveraging sophisticated invoice-themed lures and Microsoft's OAuth 2.0 device authorization flow to steal Microsoft 365 credentials and bypass multi-factor authentication.
A major US automotive manufacturer significantly enhanced its security operations by implementing behavioral sandboxing and threat intelligence, doubling its Security Operations Center (SOC) triage speed and reducing Mean Time To Detect (MTTD) to just 20 seconds.
Key findings • Six ImageMagick vulnerabilities disclosed on July 1, 2026, impacting various versions. • Vulnerabilities include memory leaks, denial-of-service, information disclosure, and ar…
Key findings • 25 Chromium vulnerabilities disclosed on July 1, 2026, all patched in version 150.0.7871.47. • Flaws include use-after-free, type confusion, and inappropriate implementation ac…
Key findings • 25 vulnerabilities disclosed together for Chrome on Android on July 1, 2026. • All flaws are patched in version 150.0.7871.47. • Vulnerabilities include use-after-free, ins…
Key findings • 25 vulnerabilities disclosed for Chrome for iOS on July 1, 2026, all fixed in version 150.0.7871.47. • Flaws include use-after-free, insufficient validation, and inappropriate …
The FTC has fined Amazon $2.25 million for failing to provide transaction records to identity theft victims, hindering their ability to dispute fraudulent charges and recover funds.
OpenMatter Network introduces a new platform using cryptographic verification to ensure auditable data usage, computation execution, and AI agent behavior, addressing trust issues in distributed environments.
Researchers have discovered a method to circumvent ChatGPT's safety filters, enabling the generation of disturbing violent and sexual imagery, highlighting persistent challenges in AI content moderation.
Anthropic's latest AI model, Claude Sonnet 5, features improved reasoning and tool-use capabilities with integrated safeguards against malicious cyber applications.
Adobe has released urgent patches for seven maximum-severity vulnerabilities affecting its ColdFusion and Campaign Classic platforms, with some flaws already being targeted by exploits.