I would say it's almost never appropriate unless you're directly quoting source texts.
If you are talking about a historical event from Egypt in 500 BCE, you would say "this happened in 500 BCE", not "this happened in year 26 of the 27th dynasty".
The distinction is important in the context of doing historical research, but it is no more appropriate to write dates in secondary sources using the Julian calendar than it would be to try to explain Egyptian history by spouting hieroglyphics at people.
Yeah, but how often are we talking about exactly specific dates without referring to specific historical texts?
The Magna Carta was sealed on June 15, 1215 for the Julian calendar. But if you want the actual anniversary, it was sealed on June 22, 1215 using the proleptic Gregorian calendar. Do you want to know the date people remember, ir the date of the true anniversary?
The coronation of Henry VIII is listed almost everywhere you look as June 24, 1509. Except that's also Julian. The date of the Gregorian calendar is July 4, 1509.
We don't remember the proleptic dates. We remember the dates that were then in use.
And if you say you're just always going to use the contemporary recorded date, that's fine. Until you get to dates like February 29, 1700. What do you put in your system then?