The level I got stuck on was after learning about scaling vertical/horizontal only, there's a level where you go under a roof and past two large boxes, then have to go up a very tall and narrow chimney-type space, with three boxes. I managed to turn them into steps once in order to get up, only to fall into an inescapable pit on the other side (I didn't extend the pipe far enough to make it to the door. Since I was stuck here, I reset and tried to get up the chimney a few more times, but couldn't make it work a second time. Part of it was the narrowness of the space (Purgob would get stuck) and part of it was the difficulty of precisely selecting which block to scale and how to scale it. The controls of QE/JK worked fine, but weren't the most intuitive for my brain to translate to take/give scale, and up/down. I'm not sure what would be better for this tbh. Maybe I needed to play a bit more for my brain to map the controls to what I wanted in my head, lol. On this particular level, at the very least that pit at the end should have some way to get out of it, especially because (if memory serves) you can't see the other side immediately after climbing out of the chimney, so there's no way for the player to tell if they'd get stuck.
Side-note on game design broadly: I agree against skipping in puzzle games overall. That being said though, I love puzzle games which give you a choice of levels to do and force you to do some, but not all in order to unlock the next sets of levels (with some exceptions). Great examples of this (and great puzzle games) include Baba is You, Bonfire Peaks, Pipe Push Paradise, and A Monster's Expedition (I recommend all these games, especially Baba is You). I find this feature empowers puzzle designers to create much harder puzzles without alienating a more casual audience. I for one love super hard puzzles, but I like to be able to bounce around between different ones while I process the challenging ones.