The snorkel library versions through 0.10.0 contain an insecure deserialization vulnerability (CWE-502) in the Trainer.load() method. This method loads model checkpoint files using torch.load() without the weights_only=True parameter, which restricts deserialization to tensor weights only. Because the default behavior permits deserialization of arbitrary Python objects via Pickle, an attacker can craft a malicious model file that, when loaded, results in arbitrary code execution on the victim system. The vulnerability is due to unsafe deserialization practices in the model loading process.

An attacker who can supply a malicious model checkpoint file to the vulnerable Trainer.load() method may achieve arbitrary code execution on the system running the snorkel library. This could lead to full compromise of the affected environment. There are no reports of active exploitation in the wild at this time.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. Until a fix is available, users should avoid loading untrusted or unauthenticated model checkpoint files with the vulnerable Trainer.load() method. Consider manually verifying or restricting model file sources and monitor for vendor updates addressing this issue.