ai-for-science-deepfri
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, prompt injections, or persistence mechanisms were detected in the skill instructions or scripts.
- [EXTERNAL_DOWNLOADS]: The skill downloads source code and pre-trained model weights from the Flatiron Institute's official GitHub and domain. The Flatiron Institute is a well-known scientific research organization, and these downloads are expected for the skill's primary purpose.
- [DYNAMIC_EXECUTION]: Static analysis flagged an
eval()call inscripts/torch_predictor.py. Manual inspection confirms this is the standard PyTorchmodel.eval()method, which is used to set the neural network to evaluation mode and does not represent an arbitrary code execution risk.
Audit Metadata