When using shotcut in hierarchical mode, meaning dragging in mlt files in a top-level project, allows combining of individual edited videos. When continuing editing on the underlying videos this can have influence on the length of those videos. In the top level project the length is not yet automatically adapted. Manual checking is required to prevent frozen frames in the output of the top level due to shortening of underlying videos.
It would make great sense to not “remember” the original size of the video when inserting into the top level, but add some dynamic snapping point method that allows to let this top level project scale with underlying edits.