GM moves and principles and, just in general the GM section of books are always the most ageda inducing section of designing a book for me. What are ya'lls takes on them, and how do you make them more compelling than just "run the dang game"
While I'm still extremely new to designing games, the thing that I find important is that the principles and moves are evocative and help the GM with keeping the game on track within whatever fiction the game is meant to go for. I was working a bit on my own game that had a very anime, specifically soul eater vibe, with a dash of Steven Universe and I'm still pretty proud of "Hint at future fights, and future friends" as a move, since I think it just really straightforward to read for the GM but also really just tells them, what the game is going to be about!
First off, co-signing what SnowyStrix said. Principles and moves are one of your key chances to inform the MC what your game is / is about / feels like in practical terms rather than aspirational ones. Try to get a clear sense of what (variously) genre, tone, mood, arc of play, feeling, etc you're aiming for and write to communicate that.
For principles, my take is this: different principles address different parts of MCing the game, and it's good to be clear which are for which as you're writing. Some principles tell you how to run the game (address yourself to the characters, not the players; make your move, but misdirect; sometimes, disclaim decision making). Some principles signpost the feeling/tone you should aim for (barf forth apocalyptica, look through crosshairs, respond with fuckery and intermittent rewards). Some principles tell you what kinds of things to put in the game (name everyone, make everyone human; ask provocative questions and build on the answers).
Most good principles, to my mind, do more than one thing at once.
Like, look through crosshairs tells you all at once that Apocalypse World is dangerous and should feel fraught and ripe with the potential of loss, and tells you that it's not your fucking job to be precious and protect people (NPCs or otherwise), and tells you when someone or something comes on screen to "consider first killing it, overthrowing it, burning it down, blowing it up, or burying it in the poisoned ground" (thanks Vincent).
I think one of the things that makes for strong, memorable, actionable principles is that multiplicity. Come up with statements that give the GM all at once a cluster of good advice.
Being poetic or metaphorical is good, if your use of those techniques is strong and clear. Principles should start with an active, actionable verb. And at least a couple of them should point pretty directly to the themes of / central assertions about the world made by the game (in AW for my money this is look through crosshairs & name everyone, make everyone human).
Also, principles are a key place to signal the voice of your game. There's a world of difference communicated by the difference between look through crosshairs (Apocalypse World), think dangerous (Dungeon World), show them the barrel of a gun (the Sprawl), hold a sword to their throats (To Tread the Spiral Path, my game), and make death common and make it matter (All Things Under Heaven, another of my games), even though we're all gesturing to a similar direction.
For moves: use an actionable verb, and be as straightforward and plain-spoken as you can. This probably isn't the place for poetry.
Dungeon World and Apocalypse World both have a section that's like (paraphrased) "these moves aren't jargon, they mean what they mean, just do the thing", but if we're real they both can get jargon-y. Inflict harm (as established) and announce off-screen badness feel intimidating and a little obscure; hurt them and show distant trouble do the same work and feel (at least to me) a lot more accessible as tools to reach for and immediately actionable to use.
All of this is, of course, just my opinion, but I hope it's helpful. At this point I've got like, 3 (4?) full-scale AW-adjacent games in the works and the MC stuff (agenda, principles, moves) feels both more intuitive to tackle and more rewarding to write with each new game.
I think what makes GM advice, and GM principles and moves feel so difficult to put together sometimes is that its hard to Articulate the ways in which you as the designer feel the game should be played into instructions that feel specific enough that they will always work while not being too specific and restrictive, and that properly represent the feeling you want to replicate.
Something that I think works well for thinking about principles, moves, and specific instructions is to think about what you would do as a GM in a variety of situations, test play the game a few times if you can, and create a list of what you did/would do then work backwards from there. Think about what you did, figure out why you did it (i.e. if you describe "the dragon as it emerged from the lake in the distances red scales glittering as it launched into the air in a spray of mist" did you do it to " Show them a dangerous creature nearby" or to "show them something fantastical in the world") and build a list. Once you're done you should have a list of things that sound like moves you made, go through and clean them up, see if any are similar or can work the same, and add some flavor to any that sound particularly boring (i.e. "Make an enemy appear" becomes "Have a monster burst forth!")
I think having a strong core and mood to the game helps a lot. The more specific and focused the game is, the easier it is to give advice about how to run it. PbtA is usually pretty good for this, seeing as most examples are trying to simulate a certain genre. For myself, based on my own experiences GMing, my primary focus is to keep things moving. Having a game stall out or get stuck in combat or whatever for an hour just isn't fun for me. It's just exhausting. Maintaining momentum and energy and keeping up that sense of progression is my number one priority. So everything is kind of built around that, I have advice for dealing with a string of partial successes, skipping ahead if a scene is dragging, introducing wrinkles or changing things up if a scene is flagging. I also include a lot of play examples, because I think they really help illustrate how mechanics can work when you're actually playing. And of course I include a bunch of caveats--this is how I play this game, but don't take it to be prescriptive. If something isn't working, drop it. If you want the game to be different, change it. If everyone's having fun and excited about what might happen next, it's all good.