Security context

What an agent needs to avoid regressing past fixes and find the next vuln in this repo.

huggingface/pytorch-image-models

              main @ e5e4dfa
            

Fixes

CVEs

CRITICAL

Peak severity

100.0%

Coverage

Highlights

Insecure Deserialization: 1 prior fix. Scrutinize any change in this area.

.github/workflows/upload_pr_documentation.yml: most-fixed (1 issue). Treat as high-risk during review.

1 high-severity fix in this history; regressions here are high-impact.

Recurring patterns

The bug types that recur here, drawn from past fixes, not open vulnerabilities.

Insecure Deserialization: The use of standard python pickle loading for model checkpoints and weights allowed arbitrary code execution. Mitigation requires restricting the unpickling environment or enforcing PyTorch's weights_only=True.

Supply Chain Vulnerability: Referencing mutable GitHub Action branches (like @main) exposes the build and release pipeline to upstream compromises. Pinning actions to immutable commit SHAs is required to secure the pipeline.