MONAI is vulnerable to OS command injection in monai/apps/nnunet/nnunetv2_runner.py. The runner concatenates user-controlled values from a configuration YAML (dataset_name_or_id) or CLI kwargs into a command string without quoting or validation, and the string is passed to subprocess with shell=True, so shell metacharacters are interpreted and arbitrary commands are executed. A victim need only load a crafted YAML in code that launches training or validation jobs. Fixed in 1.6.0, which validates dataset_name_or_id against a strict regex and builds the command as an argument list.