Ludwig framework versions through 0.10.4 are vulnerable to insecure deserialization (CWE-502) in the model serving component. The vulnerability arises because the model server loads model weight files using torch.load() without enabling the weights_only=True parameter, which restricts deserialization to tensor weights only. This default behavior allows arbitrary Python objects to be deserialized via pickle, enabling an attacker to craft a malicious PyTorch model file that can execute arbitrary code on the system hosting the Ludwig model server.

Successful exploitation of this vulnerability can lead to arbitrary code execution on the system running the Ludwig model server. This could allow an attacker to execute malicious commands or code with the privileges of the Ludwig process, potentially compromising the host system. There are no known exploits in the wild at this time.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until an official fix is available, users should avoid loading untrusted or unauthenticated PyTorch model files with Ludwig's model serving component. Monitoring vendor channels for updates is recommended.