Cybersecurity perspectives
from the frontline.

Prompt Injection Is a Permissions Problem Two national agencies say the model can't be fixed. Google says attackers aren't waiting. Here's w...

In July 2026, two of the most capable AI models in the world broke out of a locked-down test environment, found their way onto the open inte...

Two independent research papers published this year confirm what the security community has been anticipating: self-replicating AI malware i...

Three Signals, One Problem: Speed Is Outrunning Australian Cyber Controls A pattern reading of three recent stories, and what Australian sec...

A nine-year-old logic flaw hiding in the Linux kernel's cryptographic subsystem has surfaced as one of the most impactful privilege escalati...

Claude Mythos Vendor Breach: Implications for Enterprise Exposure Management A review of the reported Project Glasswing vendor breach and wh...

Two of the largest AI developers — OpenAI and Anthropic — both expanded their security-focused AI capabilities this week. For Australian sec...

Traditional penetration testing has a coverage problem. A typical web application assessment covers somewhere between 20 and 40 per cent of ...

Penetration testing has not changed much in its fundamentals over the past two decades. A skilled consultant, a defined scope, a time-boxed ...