If I do a selective file recovery, I set the task to append ".bad" to files with errors. I normally leave those files where they are and explain to the customer how to find them. I also quote the number of files & folders recovered and their total size, and add "... of which xxx files (yyyGB) were corrupted/unreadable." If I follow best practice and do a drive image (rather than selective file recovery) and recover the files from the image, there is no way to know which of the recovered files are corrupted without opening/playing/running them.
That was before got Corrupted File Finder. It can scan the recovered files and flag/separate those that are corrupted, based on them containing a sector full of the fill character. There may still be corrupted files that don't meet that definition and some files that do but are not corrupted. It's not infallible but is an improvement.
You could always do the drive image first and then run a selective file recovery task against the patient. That would get you a more accurate list of "bad" files but still allow further recovery from the image, e.g., of lost or deleted files.