TechNadu

Add as preferred source on Google Aditi Bhatnagar, CEO and Founder of Off-Grid Security, takes us through how AI agents move from spotting suspicious code to proving an exploitable attack path. Which parts of vulnerability hunting have proved difficult to hand over to an AI agent? Where does human judgment remain essential? An agent can tell you what's possible. A human decides what matters.

Students looking for textbooks got a cybercrime group's calling card instead. Amsterdam-based academic publisher Elsevier confirmed it was briefly compromised following an attack on Monday, September 21, 2026, after visitors to select platforms were redirected to a LAPSUS$ leak page. The problem surfaced on September 22, when a self-described nursing student posted a screenshot of the LAPSUS$ leak site on Reddit. They had been trying to access homework and textbooks.

Threat actors began exploiting a critical WordPress vulnerability within hours of its public disclosure. Tracked as CVE-2026-87902 (CVSS score: 9.2), the flaw lets an unauthenticated attacker achieve remote code execution (RCE) under specific server and theme conditions. The first attempt was logged on September 22, 2026, at 11:49 a.m. UTC, the same day patches were released.

Australia says an AI agent built by OpenAI breached a government health data portal in June, potentially the first known case of an AI agent hacking a government website. Prime Minister Anthony Albanese said the agent got into the medical statistics portal of Medicare, Australia's universal health insurance program, while researching public medical spending. Albanese spoke to reporters on Wednesday in New York, where he is attending the UN General Assembly.

ShinyHunters says it is holding FBI personnel data hostage, and a Reuters review shows the sample it released goes well beyond routine HR records. ShinyHunters announced Tuesday that it breached the bureau and took data on a large number of current and former employees. The group says it will not return the data until the FBI rescinds an unflattering statement about it issued in May.

A newly introduced bill in the US House of Representatives could require major VPN providers to block access to websites accused of being foreign piracy platforms. The American Copyright Protection Act (ACPA), introduced as H.R. 10364 by Representative Darrell Issa (R-CA) on September 16, 2026, would create a faster legal process for copyright holders seeking court orders against sites accused of copyright infringement. The bill is listed in the official US government record on GovInfo.

Add as preferred source on Google Zebulon “Ziggy” Griggs, Founder and CEO at Zigabyte, discusses how organizations can govern AI agents across payment and IT support workflows to prevent unauthorized actions and permissions from remaining active after they are not needed. A finance agent reads an invoice, checks a vendor record, and initiates payment. What existing identity and authorization checks govern these steps? Where could they fail? What should be added before money moves?

An Armenian national extradited from Ukraine has been sentenced to 24 months in federal prison for his role in Ryuk ransomware attacks and an extortion conspiracy that hit companies across the United States, including in Oregon. Karen Vardanyan, 35, was also ordered to pay $1,219,106.00 in restitution to his victims. U.S. Attorney Scott E. Bradford for the District of Oregon announced the sentence on September 22, 2026.

Microsoft has disrupted EvilTokens, an AI-powered cybercrime platform, working with U.K. police to arrest two suspected operators. U.K. officers executed warrants at two locations last Friday and arrested two men on suspicion of making articles for use in fraud and money laundering. Both were released on bail while the investigation continues. Microsoft reported the platform to London's Metropolitan Police Service in August.

CLOSEDQUORUM is what researchers call the first publicly documented Windows implant to hand tactical command-and-control (C2) decisions to artificial intelligence (AI) rather than a human operator. Cisco Talos detailed the malware in a new report, as the inaugural finding from Talos' new CAIRN project, a toolkit built to track AI-integrated malware.