SSU and FBI exposed a sustained Russian intelligence campaign using fraudulent support texts to compromise messaging accounts of Ukrainian officials, military, politicians, and European and U.S. targets.
OpenAI released three GPT-5.6 variants—Sol, Terra, Luna—in restricted preview to select companies under a U.S. government engagement, with Sol carrying the company's toughest cybersecurity guardrails to date.
Ransomware and data-theft attacks routed through third-party vendors are accelerating across education institutions, exposing student records at scale. The sector's weak vendor-risk frameworks make it a repeatable soft target as threat actors pivot from direct intrusion to supply-chain leverage.
Threat actors are mass-producing pig-butchering investment scam sites using templates built on the legitimate Chinese DCloud Uni-App framework, scaling fraud infrastructure to 200,000 domains.
A proof-of-concept attack embeds a malicious payload inside a superficially clean GitHub repository that bypasses security scanners, AI reviewers, and human auditors, executing when an agentic coding tool clones and sets up the project.