CyberRota Analysis
AI-GeneratedThe vulnerability in SGLang allows for remote code execution due to insecure deserialization of .bin files when loading model weights from a HuggingFace repository. This critical flaw, triggered during the /update_weights_from_disk process, poses significant risks to any system utilizing this functionality, potentially allowing attackers to execute arbitrary code. Organizations using SGLang should prioritize immediate remediation to mitigate the risk of exploitation.
Public Exploit Signal
A public exploit, PoC, GitHub repository or Metasploit reference was detected for this CVE.
Note: these links are listed for security research and verification purposes only.
Original NVD Description
SGLang contains a RCE vulnerability when attempting to load model weights from a HuggingFace repository, specifically within the /update_weights_from_disk, where torch.load(..., weights_only=False) fallback enables pickle deserialization of .bin files.
Related CVEs
Other vulnerabilities affecting the same vendor(s)