Stories cluster related articles into a single narrative, linked to the underlying CVEs and affected products. 8,564 stories synthesized.
CISA's red team exercises revealed stark differences in cybersecurity defenses, with one organization falling victim to undetected domain compromise while another rapidly detected and contained threats.
The ToxNetV2 Linux botnet is evolving by incorporating NVIDIA's AI services to assist threat actors in generating shell commands and remote SSH attack actions, marking a significant step towards AI-assisted live offensive operations.
Fashion retailer ASOS has alerted customers to unauthorized access to their accounts, likely resulting from credential stuffing attacks that leveraged externally sourced login information.
A coordinated cyberattack leveraging eight AI agents successfully breached Asian government systems, compromised 85 accounts, and exfiltrated over 2,500 personnel records.
A flaw in NVIDIA NemoClaw enables malicious webpages to unauthenticatedly control local Ollama instances, potentially injecting hidden instructions into AI models.
A critical flaw in Tata's B2B procurement platform, Nexarc, allowed attackers to take over accounts using only a registered phone number by exploiting how one-time passwords (OTPs) were handled.
INTERPOL's Operation Jackal IV, a multi-national effort, resulted in 58 arrests and identified 263 suspects linked to West African organized crime groups, highlighting evolving criminal tactics.
Microsoft's August 2026 .NET Framework cumulative updates are causing System.IO.FileFormatException errors in WPF applications when generating PDFs/XPS or printing, impacting multiple Windows versions.
WhatsApp is rolling out enhanced security features, including support for multiple passkeys and a strengthened two-step verification process, to improve user account protection.
Multiple vulnerabilities in Zscaler Client Connector (ZCC), including CVE-2026-59568, enable unauthenticated remote code execution with low complexity.
A critical vulnerability in Marimo notebook software, CVE-2026-75149, allows attackers to execute arbitrary commands by crafting malicious notebooks that are opened in edit mode.
Broadcom has disclosed 91 new vulnerabilities impacting the Spring Framework and its related projects, affecting an estimated 209,569 software components across the open-source ecosystem.
A campaign has been discovered leveraging 24 npm packages to host fake Cloudflare CAPTCHA pages, redirecting users to phishing sites.
A large-scale distributed denial-of-service (DDoS) attack has crippled Norwegian public services for over a day, impacting critical infrastructure like the national identity gateway.
A novel prompt injection technique, Cryptographic Context Injection, hides malicious commands within encrypted data, tricking AI assistants into executing them after decryption.
Two new remote access trojans, E4del and PINHOLE, are employing a novel technique using FTP banners as dead drop resolvers to hide command and control infrastructure.
A critical flaw in Provenance Blockchain's marker module allowed any user to gain administrative privileges over 82 live financial asset markers, potentially enabling token minting or escrowed asset theft.
Australia has issued a warning about the active exploitation of a critical vulnerability affecting TeamCity servers, following a similar alert from the US government.
Take-Two Interactive is seeking user data from Microsoft and Discord to identify the perpetrator behind the Grand Theft Auto 6 leak, potentially exposing thousands of users.
Despite growing concerns, the vast majority of AI-authored malware samples are confined to research labs and sandboxes, with existing security tools effectively detecting and blocking those that attempt to reach production environments.
Cisco Talos warns that restrictive guardrails on cloud-hosted AI models create a 'safety penalty' for cybersecurity defenders, while adversaries exploit unconstrained models.
The EvilTokens service uses AI to analyze stolen Microsoft 365 session data, identifying key contacts and financial flows to craft highly convincing BEC scams.
Attackers are using fake OpenAI Codex and Anthropic Claude Code download ads on Google Search to trick Mac developers into running malware via a 'ClickFix' technique.
Taiwanese prosecutors have charged nine individuals, including employees from Nvidia and Super Micro, for illegally exporting advanced AI servers and semiconductors to mainland China.