Ken Underhill

OpenAI’s GPT-5–powered Aardvark autonomously finds and fixes software vulnerabilities, redefining AI-driven cybersecurity.

Researchers discovered agent session smuggling, a new attack where rogue AI agents secretly inject commands to deceive and manipulate other agents.

Progress patches a MOVEit Transfer flaw letting attackers exhaust resources and cause denial-of-service without authentication.

Threat actors misuse AzureHound to map Azure and Entra ID, turning a security tool into a powerful cloud reconnaissance weapon.

Cybercriminals are turning AdaptixC2, an open-source security tool, into a weapon for ransomware attacks.

Chinese-linked group UNC6384 targets European diplomats with a Windows shortcut exploit to deploy PlugX malware.

A new exploit called Brash can crash Chromium browsers in seconds by overloading the tab title API.

Hidden npm malware steals developer credentials, exposing major software supply chain risks in the open-source ecosystem.

Hackers breached Canadian water, energy, and farm systems, prompting national warnings to secure industrial control networks.

A public exploit for a critical BIND 9 flaw renews DNS cache-poisoning risk, enabling forged records and traffic redirection.

Sandworm targets Ukrainian networks using stealthy, low-malware attacks that exploit legitimate Windows tools to evade detection.

A Windows flaw in the Narrator tool enables DLL hijacks and persistence. Learn how attackers exploit it and how to harden systems.