I must admit, the GUI doesn't speak for itself but it still should be possible to figure things out.
As for the object interaction, can't you do it so that the user manually makes the path (including interactions) and that the script repeats it?
Sent from my phone using shitty auto-corrections.