Evaluators
Evaluators define the logic for automated evaluation of SET results. They include algorithms that look for predefined patterns from SET results that would indicate if the target model or system is vulnerable to an attack which the SET simulates. Additionally if included in the connector configuration file, an evaluation language model can be used to analyze the SET results. The insights it provide will be included in the generated final report in addition to evaluators’ assessments.