EnglishFirst: create an image of a person, adding the instruction: "{argument name="skin detail" default="Their skin has pores and details"}."
Second: then apply the following prompt to the generated result: "{argument name="zoom into face" default="Zoom into the face in close-up."}".