Small world: The revitalization of small AI models for cybersecurity
Sophos X-Ops explores why larger isn’t always better when it comes to solving security challenges with AI
Cybersecurity never slows down. Every week brings new threats, new vulnerabilities, and new lessons for defenders. For security and IT teams, the challenge is not just keeping up with the news—it’s knowing which risks matter most right now. That’s what this digest is here for: a clear, simple briefing to help you focus where it…
Cybersecurity researchers have discovered three malicious Go modules that include obfuscated code to fetch next-stage payloads that can irrevocably overwrite a Linux system’s primary disk and render it unbootable. The names of the packages are listed below – github[.]com/truthfulpharm/prototransform github[.]com/blankloggia/go-mcp github[.]com/steelpoor/tlsproxy “Despite appearing legitimate,
He called himself ALTDOS when he first contacted DataBreaches in 2020. In 2021 he started contacting this site as DESORDEN. Then in 2024, he contacted this site as GHOSTR, and more recently, as 0mid16B. Under each new moniker, he denied being the individual DataBreaches knew under previous monikers, even though based on his targets, his…
Business Recorder reports: Pakistan Petroleum Limited (PPL), an oil and gas exploration firm, has reported a ransomware attack on parts of its IT infrastructure, detected on August 6, 2025, but said the incident was swiftly contained with no compromise of critical systems or sensitive data. The E&P disclosed the development in its notice to the…
Cybersecurity researchers have disclosed details of a new campaign that uses WhatsApp as a distribution vector for a Windows banking trojan called Astaroth in attacks targeting Brazil. The campaign has been codenamed Boto Cor-de-Rosa by Acronis Threat Research Unit. “The malware retrieves the victim’s WhatsApp contact list and automatically sends malicious messages to each contact…
Cybersecurity researchers have disclosed a critical security vulnerability in Ollama that, if successfully exploited, could allow a remote, unauthenticated attacker to leak its entire process memory. The out-of-bounds read flaw, which likely impacts over 300,000 servers globally, is tracked as CVE-2026-7482 (CVSS score: 9.1). It has been codenamed Bleeding Llama by Cyera. Ollama is a