pytorch-patterns

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFE

Full Analysis

[SAFE]: The skill provides idiomatic PyTorch patterns for common tasks like model definition, training loops, and data loading.
[SAFE]: It emphasizes security by recommending weights_only=True for torch.load(), which protects against arbitrary code execution vulnerabilities common in Python's pickle module.
[SAFE]: No malicious behavior, unauthorized network access, or sensitive data exposure was found in the provided instructions.

Audit Metadata

Risk Level

SAFE

Analyzed

Sep 14, 2026, 07:59 PM