torch.load when serving Ludwig models (ludwig serve)MITRE service request: 1988584
Status: RESERVED (pending a qualifying public reference per CNA Rules §5.3).
The Ludwig framework thru 0.10.4 is vulnerable to insecure deserialization (CWE-502) in its model serving component. When starting a model server with the ludwig serve command, the framework loads model weight files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the pickle module. An attacker can exploit this by providing a maliciously crafted PyTorch model file, leading to arbitrary code execution on the system hosting the Ludwig model server.
ludwig serve -m /path/to/model loads weights with unsafe torch.load. Hosting a malicious .pt archive leads to instant RCE on the inference host during startup.
00c51e0a286c3fa399a07a550e48d0f3deadc57d).ludwig serve model bundle path.torch.load without weights_only=True.High for multi-tenant inference gateways.