1 paper
Daniel Begimher, Cristian Leo, Jack Huang +2
We present SIR-Bench, a benchmark of 794 test cases for evaluating autonomous security incident response agents that distinguishes genuine forensic investigation from alert parroti…