To me, it depends. I think there are junk profiles -- where the profile is so messed up and conflicted from multiple errors in copying, accidentally attaching wrong people in wrong places, and even just plain bad assumptions while researching that it is hard to view as representing an actual individual.
If I can find evidence that the person existed at all, I don’t mind adding the lipstick. For instance, if the profile Abigail Jones was clearly created as a mother for John Smith, and that John Smith is documented somewhere, then regardless of whether his mother was an Abigail or a Jones, she existed. The bio can be sufficiently weakened to say everything lower down on the profile is dubious, at best.
But in the back of my mind, I have been trying to figure out what to do with the following three profiles:
https://www.wikitree.com/wiki/Wallis-661
https://www.wikitree.com/wiki/Wells-5687
https://www.wikitree.com/wiki/Welles-94
I guess at least one was supposed to be the father of
https://www.wikitree.com/wiki/Welles-29
But that profile already has a father, and it’s not Thomas. Is there some other Thomas producing this data?
My interest was drawn by the wife Abigail Thurber attached to the first profile I linked. But I don’t know who that Thomas was supposed to be, and I think Abigail is just a non-person. Not “uncertain existence”, because to me that connotes that there was at least a mythical person in oral history or something but who may have been fabricated into the story. I think Abigail Thurber was more likely just the result of probably several mistakes as described above. IMO it would be easier just to delete the profile. I think lipstick in this example would be harmful, for the reason you state.