NVIDIA NeMo Framework: Malicious input causes improper control of code generation, reaching code execution

Impact

Malicious input causes improper control of code generation, reaching code execution. In an AI datacenter this is the model-and-data supply chain problem: the code runs with whatever the training or inference job holds, which is usually a GPU, a service account, and mounted object storage credentials.

Who can reach it

Requires the job to load an attacker-influenced artifact - a checkpoint, .nemo file, config, tokenizer or dataset. Any pipeline that pulls from a public model hub, a customer bucket, or a tenant-supplied path is in scope.

What to do

Bump the package to the fixed version in bulletin 5718 and rebuild every training/inference image that embeds it. Cost: image rebuild and job restart; no host driver or firmware change. The durable control is refusing to deserialize untrusted checkpoints at all - prefer safetensors-style formats and treat pickle-bearing artifacts as executable code.

References