How vulnerable is my policy? Adversarial attacks on modern behavior cloning policies