I've been worried about a lot of your feedback for a while, particularly the box spam- I used it during the jam because it's a quick way to make puzzles that feign actual difficulty, and it managed to fool a couple people then, but now that I have more levels and time to work on it people are starting to catch on XD.
I'll probably end up phasing out crates altogether by the end of Day 2, but I'm keeping them around in the earlier levels, both to introduce newcomers to block-pushing games and to strengthen the theme of a cluttered workspace run by a grumpy master who is too experienced and enlightened to bother with material matters like tidying up. With more time I'll be able to tighten up the rooms and practice using the key mechanics more - the level you mentioned at the end of Day 1 was my favorite to design and definitely my best puzzle because it focused on keys and not crates.
Red doors were just completed, so I haven't come up with a lot of good puzzle concepts yet, but I'll definitely go a lot more in-depth with them before moving on to my plans for Day 3 and Day 4. The 5 door level was designed as a tutorial for the dungeon locks (demonstrating how they can allow you to move fragments without permanently affixing them to your handle) and has some interesting pit puzzles, but it can absolutely be made way smaller and less painful to push through.
Not reducing the weld cooldown- in fact I've been considering increasing it to make certain welds possible, but reducing it even by one turn breaks almost all my puzzles as is. Even with smaller levels I don't think it will be necessary.
Thank you for your very detailed criticism and advice! This may be the longest reply I've received to date and will help me a lot with the next few updates.