1 paper · 1 filter
Pedro Conde, Henrique Branquinho, Valerio Mazzone +3
AI pentesting agents are increasingly credible as offensive security systems, but current benchmarks still provide limited guidance on which will perform best in real-world targets…