This is actually sort of like how I mod, though when I make a mod or a weapon I've usually written down the idea and fleshed it out a bit first (mostly because I want to get the idea written down before I forget stuff).
However, since I almost never end up with exactly what I set out to make, I often deviate heavily or throw ideas away if they don't work in practice, and the best ideas often come in the process of implementing the original idea, I don't like to plan too much ahead.
For example, with weapons I almost never write down exact damage amounts. Instead I'll say "Blaster: moderate fire rate, fires bursts of glowey energy, each shot deals low-moderate damage. Cheap to buy, good early weapon". No exact details or anything (22 damage per shot, 0.6 fire rate, etc), just the general idea to remind me of what what in my head at the time, and some thoughts on how it fits into the weapon hierarchy etc. Diving right in with a vague or partially fleshed-out idea often works best for me, imo. Often playing with the idea and testing it out as I go will produce much more realistic plans and details than what I can come up with on paper in advance.
Same with mapping. I have an idea of what some of the rooms will look like and how the map might play, and then I want to hop into the software and start pick'n at it until I have a map.
