This early in the game, it's tempting to start coding up a whole bunch of skills and crafting recipes, but it's sort of a mistake. This phase is what they call building a vertical slice: a small gameplay "area" where you can test out everything the game has to offer. Until you get your vertical slice built and tested, it can be a waste of time to start piling on the skills and recipes and icons. At least for me.
However, at some point you do need to start making concrete data, data that can form the foundation of the game. When is that? I couldn't say, honestly. It's something I've always had a hard time knowing: when to transition from building "test" this and "experimental" that, over to constructing the final (or at least semi-final) game data. When do I stop building throwaway goblin flank recipes that flag you as a cannibal, and start on the "real" stuff? (That's not to say the Roasted Goblin Flank won't make the cut; it's still early to rule that out).
There is just still so much I have to do before I can really worry about that, though. I mean, I don't even have a proper AI yet. The enemy mobs are still just running around on a TestRandomWander controller. You might as well call them Ambulatory Goblin Flanks, for all the real challenge they pose for a hungry Goblinson Crusoe, well-stocked with tinder for the casting of fireballs.
I did work today on the stuff to generate the spell description text. It was hard-coded before; now it uses the spell name, description, and required-resources data to build it. I need to add color-coding at some point, and that's going to complicate it a little bit. I did change the spellbook population code to adjust the spacing, in order to allow for longer descriptions than can fit in a 64-unit tall box.