There is a similar passage (in terms of the use of tenses) in the Nisus & Euryalus story in book 9:
416-419
Diversi circumspiciunt. Hoc acrior idem
ecce aliud summa telum librabat ab aure.
Dum trepidant, iit hasta Tago per tempus utrumque
stridens traiectoque haesit tepefacta cerebro.
The imperfect describes the action that is going on in the background here while the perfect tense foregrounds the main action (and the present in 'dum trepida(n)t' seems to underline that even more).
I think he was trying to do the same thing in the passage you quoted and simply ignored the technicalities of reported speech. It wouldn't be the only time a Latin writer doesn't play by the rules of the grammar book
(I know that Ovid sometimes puts indirect questions in the indicative without any necessity)
If it had been me writing these lines, I would probably have started with the
conscendebat equos part and then I would have gone on and tried to put the main clause in metres around it ... then I would have got annoyed because it doesn't fit as I need a short syllable at the end somewhere ... then I would have turned it into indirect speech to gain that short syllable from the infinitive (
rapuisse), at which stage I would have forgotten that I would need to adjust the subordinate clause as well
(and the adjustment wouldn't be super-easy as you can't just replace
conscendebat with
conscenderet) ... But I don't want to imply that Vergil necessarily worked that way
... Or maybe he thought, "Let's just leave it at that for now, I'll look over the whole thing when I'm back from Greece."