Well, the first step is to run the command, and redirect the output to a file.
Then the next step is to grab specific text from the file into a second, smaller file, probably matching via a regular expression.
However I don’t think the task – identify “unlucky” plots – is a good idea… you want to remove bad plots for sure, ones that fail plot check, but don’t worry about the unlucky stuff, it is a waste of time…