In this replication package, we provided Mithra and AR-SI's implementations, the dataset of 24 bugs on ArduPilot v3.6.9 that we used for evaluation, and 2500 traces of unlabeled training data and 466 evaluation traces (233 erroneous and 233 correct traces). In addition, we provide the dataset of 153 bugs on F1tenth that we used for evaluation, and 75 traces of unlabeled training data and 522 evaluation traces (261 erroneous and 261 correct traces).