Sometimes I do work on the game, but it's pretty mentally taxing, so I usually take lots of breaks. As for the puzzle design process, Lockpick is kind of a unique case since it's primarily rooted in deductions and untangling dependency chains, but more generally, I design puzzles by trying to make the "essence" first, whether it's a new interaction or a fundamental realization about the puzzle which informs how you need to approach it. Then, I try to make a setup for that trick with whatever pieces I can think of, the simpler the better (as more complex puzzles run a greater risk of having an unintended solution). It's kind of up to the designer how to do this part, including whether or not to add bait elements which distract from the solution, which was the case for a lot of Lockpick's puzzles. That's not the only way to obscure a solution, though, and a secondary goal or element which has a different use from the player's initial obvious assumption are also good ways to do it. My first real puzzle experience was Jelly no Puzzle, which is where I picked up this design philosophy from.