pytorch-patterns
Pass
Audited by Gen Agent Trust Hub on Apr 3, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill file consists solely of documentation and static code snippets illustrating PyTorch best practices.\n- [SAFE]: No remote code execution, external script downloads, or data exfiltration patterns were identified in the analyzed content.\n- [SAFE]: The skill actively promotes security by instructing users to use
weights_only=Truewhen callingtorch.load(), which is a standard mitigation against unsafe deserialization vulnerabilities in PyTorch.
Audit Metadata