import numpy as np
import pickle
class VulnerableModel:
def __init__(self):
self.weights = np.random.randn(10)
self.bias = np.random.randn()
def predict(self, X):
# VULNERABLE: No input validation
return np.dot(X, self.weights) + self.bias
# Create and save a sample model
if __name__ == '__main__':
model = VulnerableModel()
# Save the model (vulnerable serialization)
with open('sample_model.pkl', 'wb') as f:
pickle.dump(model, f)
# Example prediction
X = np.random.randn(10)
print(f"Example prediction: {model.predict(X)}")