Please implement measures to prevent shrinkage to seeds in the event of an issue. If there is a boundary line concern, why does the "shrink to seeds" process continue? We often encounter problems where the system locks up or times out, necessitating the creation of a new version. If the system can recognize an issue during the "reconstruct from seeds" process, shouldn’t it also do the same for the "shrink to seeds" operation?