I think the program should warn the user and reject that action. Same for other user misakes.
It need be only covered in the user manual. It isn't a mistake as you put it (in fact quite the opposite it is intentional). Nor is it a bug. Look back to my post where I have indicated a flight from LIMZ to LIMF. The last wypt is the start of the STAR. This was a test fpl during early beta testing. And it does work. Both as a STAR and vectoring. Note though that the last wypt MF503 is 22nm from Caselle. The whole fpl is only 41nm!
The end of SID is 13nm from LIMZ a 6nm gap between that and the start of the STAR at MF503.
Yes, there maybe the risk that the last wypt selected as a STAR might not work. Most likely because it is too close to the destination. In which case do as Dave suggested and add another.