From elastic/elasticsearch#18201 (comment):

For scripts, we also only test a fictional search script to at least have something. But for example, we don't even have a fictional update script yet, forget about having any variety in the types of scripts we want to benchmark or anything like that.