You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Compressed Joblib ModelScan Bypass

This PoC demonstrates a ModelScan detection bypass for compressed Joblib model files.

Joblib supports compressed persistence through joblib.dump(..., compress=...) and transparently decompresses those files in joblib.load(). The compressed files in this repository contain the same pickle payload as the uncompressed control file.

Files

control_uncompressed.joblib: uncompressed control payload. ModelScan reports one CRITICAL issue.
bypass_zlib_compressed.joblib: compressed Joblib payload. ModelScan reports 0 issues.
bypass_gzip_compressed.joblib: compressed Joblib payload. ModelScan reports 0 issues.
bypass_bz2_compressed.joblib: compressed Joblib payload. ModelScan reports 0 issues.
bypass_lzma_compressed.joblib: compressed Joblib payload. ModelScan reports 0 issues.
build_and_verify.py: rebuilds and verifies the PoC files.
verify_report.json: local verification output.

Verification

Run:

python build_and_verify.py

Expected result:

joblib.load() executes the payload from all five files and writes joblib_compressed_marker.txt.
ModelScan reports CRITICAL=1 for control_uncompressed.joblib.
ModelScan reports total_issues=0 for every compressed .joblib file.

The payload is local-only and writes the marker text JOBLIB_COMPRESSED_RCE.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support