Locking it down: A new technique to prevent LLM jailbreaks
Following on from our preview, here’s the full rundown on LLM salting: a novel countermeasure against LLM jailbreaks, developed by AI researchers at Sophos X-Ops
Similar Posts
⚡ THN Weekly Recap: New Attacks, Old Tricks, Bigger Impact
Bys sCyber threats today don’t just evolve—they mutate rapidly, testing the resilience of everything from global financial systems to critical infrastructure. As cybersecurity confronts new battlegrounds—ranging from nation-state espionage and ransomware to manipulated AI chatbots—the landscape becomes increasingly complex, prompting vital questions: How secure are our cloud environments? Can our
Personal Details of Thousands of Border Patrol and ICE Agents Allegedly Leaked in Huge Data Breach
Bys sTom Latchem reports: Sensitive details of around 4,500 ICE and Border Patrol employees—including almost 2,000 agents working in frontline enforcement—have allegedly been released by a Department of Homeland Security whistleblower following last week’s fatal shooting of Renee Nicole Good. The Jan. 7 killing of the mother by ICE agent Jonathan Ross in Minneapolis, Minnesota, has……
Open VSX Bug Let Malicious VS Code Extensions Bypass Pre-Publish Security Checks
Bys sCybersecurity researchers have disclosed details of a now-patched bug impacting Open VSX’s pre-publish scanning pipeline to cause the tool to allow a malicious Microsoft Visual Studio Code (VS Code) extension to pass the vetting process and go live in the registry. “The pipeline had a single boolean return value that meant both ‘no scanners are…
The art of restraint: why not every system should be owned
Bys sJesse William McGraw writes: Greetings, reader! I want to talk about my former life as a blackhat hacker because it never makes any sense. Many of my rampant hacking campaigns focused on breaking into big servers across several industries but not on seizing the networks by the throat and forcing them to their knees. I…
AI-Generated Malicious npm Package Drains Solana Funds from 1,500+ Before Takedown
Bys sCybersecurity researchers have flagged a malicious npm package that was generated using artificial intelligence (AI) and concealed a cryptocurrency wallet drainer. The package, @kodane/patch-manager, claims to offer “advanced license validation and registry optimization utilities for high-performance Node.js applications.” It was uploaded to npm by a user named “Kodane” on July 28, 2025. The
India’s RBI Introduces Exclusive “bank.in” Domain to Combat Digital Banking Fraud
Bys sIndia’s central bank, the Reserve Bank of India (RBI), said it’s introducing an exclusive “bank.in” internet domain for banks in the country to combat digital financial fraud. “This initiative aims to reduce cyber security threats and malicious activities like phishing; and, streamline secure financial services, thereby enhancing trust in digital banking and payment services,” the…