Thanks again for providing the project and crash log.
I couldn't make the crash occur deterministically, but the crash log contained enough information to identify and harden a genuine risk. The failure happened while a queued UI callback was updating a native menu. A debugger pause can substantially lengthen that callback’s lifetime, allowing the native menu to be replaced or released before the callback resumes.
I've made a narrow patch for the next release which I hope will fix this: https://feedback.objo.dev/bug/849
If it recurs, please let me know and it would be awesome if you could provide a crash log too.
I've changed Objo so native menus are retained for the complete asynchronous operation. Queued callbacks are now rejected if the menu has subsequently been replaced or the application is shutting down, and all native resources are released exactly once.