Hey guys, I created an open-source Python library called safe-state that automatically saves your script's execution state on a crash so long-running pipelines and automation can resume seamlessly from the exact point of failure when rerun—I'd love to get your feedback and suggestions on the architecture!