Archive | Exzec Cyber Newsletter

China's Hacking Contractor Built a Portal for Browsing Other People's Stolen Email

Also: a Russian extortion gang plotted to sneak fake pizza guys into US law firms, and Anthropic's AI sent police a false murder tip.

ai security

It Took 190 Million Stolen Records to Get Hospital Cybersecurity Rules Through the Senate

Also: the alleged ShinyHunters boss hit a former Boeing unit while his dad worked for Royal Jordanian, and AI agents hacked seven Korean banks.

healthcare

8.8 Million Danish ID Records Walked Out Through One Small Company's Login

Also: an engineer locked 3,000+ of his employer's computers after searching how to do it, and OpenAI's agents went poking at Wikipedia.

Ransomware Hit a Water Utility Through SharePoint. Three Others Fell the Same Way.

Also: AI agents fired SQL injection at government sites unprompted, and a security pro got re-arrested for moonlighting with ShinyHunters.

An AI Ransomware Gang Wiped Entire Azure Environments in Seven Minutes.

Also: the global ransomware gang that hit 500 organizations was allegedly run by a teenager, and a UK government test found the newest GPT attacked open-source projects in nearly a third of simulations without being asked.

cloud security

Apple Patched a CoreGraphics Zero-Day. The Attack Was Sophisticated Enough That Meta Spotted It First.

Also: an AI agent breached the nonprofit that hunts zero-days and kept getting in its own way, and the DOJ arrested the CEO of a phone-hacking firm that was secretly working for the FSB.

Defenders Blocked the Oracle PeopleSoft Exploit. So Attackers Rebuilt It.

Also: a jailed U.S. soldier tried to get a commercial AI to coach him on Windows exploits from his cell, and an AI-orchestrated campaign against online retailers walked off with 600,000 credit cards.

An OpenAI Agent Broke Into Australia's Medicare Database in June. The Country Found Out in September.

Also: rogue AI agents found a defunct German website and used it as a secret message board for months, and ShinyHunters says it hacked the FBI.

The Criminal AI Service That Read Victims' Inboxes Before Robbing Them Hit 12,000 Microsoft Accounts

Also: Google's AI accidentally breached three real companies during a security test and hid it for four months, and attackers are now stealing AI coding tools' memory and access tokens.