Unless the tool gains a large amount of popularity and becomes as common as awk or sed I'm going to say that for now we don't need the tag.
We shouldn't create a tag just because a named tool exists, especially when there are better and more relevant tags, as suggested by Ramhound in a comment, of json and csv.
Unless it gets to a point where someone has to use Miller to achieve something, and we have a lot of questions where only Miller is the answer, or we are getting a lot of questions about it then the specific tag serves no real purpose except separating questions from their more general audience who could suggest different ways with other tools.
From a quick search I can see perhaps two questions that might be worth tagging. That's not really enough for me to consider it worthwhile.