CVE-2026-61500 is a critical authentication bypass in Rejetto HTTP File Server, discovered by Anthropic Mythos AI and exploited within 24 hours of public disclosure by a China-linked actor targeting ...

Mythos AI found a critical Rejetto HFS flaw enabling admin-session forgery and arbitrary code execution via predictable Math.random() keys.

AI-assisted research uncovered a critical Rejetto HFS flaw that enables auth bypass and remote code execution, now exploited ...

China-linked UNC3569 exploited a Sogou Input Method flaw to deploy GRAYRABBIT; Tencent fixed the issue in version 16.3.0.3498.

GLM-5.3, an open-weight model from the Chinese lab Zhipu AI, can build working cyber exploits on its own almost as well as ...

Google confirmed on September 8 that attackers were already exploiting a flaw in Chrome’s V8 engine when it shipped the fix, ...

Explore the latest news, real-world incidents, expert analysis, and trends in Vulnerability — only on The Hacker News, the ...

Zhipu's open-weight model GLM-5.3 writes cyber exploits nearly as well as Claude Mythos Preview, according to Anthropic. Its smaller Flash variant put together a reliable Chrome attack for just $20.40 ...

Anthropic says China's GLM-5.3 can find vulnerabilities and build end-to-end exploits, while its open weights make safeguards easier to modify, raising concerns over access to cyber capabilities ...

GLM-5.3 can autonomously build end-to-end cyber exploits, but unlike other frontier models, it was released without ...

Anthropic made claude.ai faster by about 3x in a two-week sprint, with Claude writing fixes across 3,000+ changes and no reported rollbacks.

Anthropic reports that Z.ai's open-weight GLM-5.3 comes close to Claude Mythos Preview on certain exploit-development tests.