研究发现GPT-5.5在网络安全能力上与备受吹捧的Mythos不相上下
英国AI安全研究所(AISI)的最新研究表明,尽管Anthropic公司曾大肆宣传其Mythos预览模型的网络安全威胁,并限制了其发布,但OpenAI的GPT-5.5在网络安全评估中的表现与其不相上下
A weaponized proof-of-concept (PoC) exploit framework dubbed “cPanelSniper” has been publicly released for CVE-2026-41940, a maximum-severity authentication bypass in cPanel & WHM that has already led to the compromise of tens of thousands of servers worldwide with attack activity traced as far back as late February 2026. CVE-2026-41940 is a critical pre-authentication flaw rooted in […]
The post cPanelSniper – PoC Exploit Disclosed for cPanel Vulnerability, 44,000 Servers Compromised appeared first on Cyber Security News.