- Introduce --all argument to run_training_test.py to execute all available scenarios sequentially. - Refactor single scenario execution into a run_single_scenario function for better modularity and error handling. - Implement a print_summary function to report results for multiple scenario runs.