1 paper
Liwei Yu, Shuo Li, Ming Zhou +2
Large Language Models (LLMs) have shown promise for automated penetration testing, yet existing end-to-end black-box evaluations are highly susceptible to error cascading: failures…