Wow, this is incredible feedback! I think you hit the nail on the head with your first two points; I've felt that grounded and aerial acceleration / deceleration has always been my Achilles' heel when it comes to making a platformer. Splitting the mid-air vectors like you mention sounds so simple in hindsight, but it had never once crossed my mind as a solution.
The lack of snappyness in the attack was definitely a product of both the jam’s time constraint and me not being the most experienced with animation (and more importantly, blending animations together). Certainly something I would have liked to circle back on had time permitted. It’s a similar story with the sprinting. We had noticed the slow startup from zero speed only after many of the level assets had already been modeled, so I was hesitant to mess with the movement much more and risk running out of time if the level design needed to be fixed to account for that sort of change.
I really appreciate these suggestions, I can see myself returning to this comment many times in the future whenever I’m starting similar projects. Thank you!