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.

ModelScan Bypass PoC — ctypes.CDLL + operator.methodcaller

Demonstrates that ctypes.CDLL combined with operator.methodcaller bypasses modelscan v0.7.6 for native code execution.

Vulnerability

ctypes.CDLL(None) loads the default C library (libc)
operator.methodcaller("system", b"cmd") calls .system(cmd) on any object
Combined: executes arbitrary shell commands via libc's system() function
Neither ctypes nor operator.methodcaller are in modelscan's blocklist

Impact

Native RCE — bypasses Python exec entirely. Calls libc system() directly via ctypes FFI. This is the most dangerous variant as it cannot be mitigated by Python-level sandboxing.

Reproduction

pip install modelscan==0.7.6
modelscan scan -p pytorch_model.bin
# Output: No issues found

python3 -c "import pickle; pickle.load(open('pytorch_model.bin','rb'))"
# Executes arbitrary shell commands via libc

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