1) Cells are put in a dish
2) Gene "editing" tech is added to the dish
3) Most, if not the vast majority of the cells die. Or this information is simply not reported, even if the authors did report it in other papers so they know it is important.
4) The surviving cells consist largely of those with "edited" genes. Control cells, who were not exposed to the "editing" conditions, also have a small percentage of "edited" genes.
It sounds to me like they are selectively killing the cells that are not already mutants. I read around 8 papers on CRISPR, none of them addressed this possibility (or mention it and cite someone who did address it).
Of course, in terms of getting the desired mutants for the clinic it works just as well. Misunderstanding why it works is not good though.
Live long enough, and we'll keep seeing amazing things!