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.

MFV PoC: typing.ForwardRef Pickle Chain (Scanner Bypass)

WARNING: This repository contains a proof-of-concept malicious model file for security research purposes only.

Vulnerability

The file malicious_model.joblib contains a pickle payload that achieves arbitrary code execution when loaded via joblib.load() or pickle.load(). The payload uses typing.ForwardRef and typing.get_type_hints to trigger compile() + eval() internally, bypassing all major model file scanners.

Reproduction

import joblib
# This will execute: touch /tmp/pwned
model = joblib.load("malicious_model.joblib")

Scanner Results

picklescan 1.0.4: 0 issues, 0 infected files
modelscan: "No issues found"
fickling: No specific dangerous import flagged

GLOBAL References (none on any denylist)

typing.ForwardRef
typing.get_type_hints
builtins.type
builtins.object

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