2 ms·Susceptibility of Pre-Trained Language Models to Handcrafted Command Injection2 points by upwardbound 4y ago