I don't negotiate with them, play mother may I or hand waive anything. They simply tell me what they are doing and I react accordingly. If they tell me that they are going to the port city of Athkatla to steal a ship, that's what they are doing. When they get there they will let me know how they go about it, and the game world will react to it.
And this is always my approach to GM-led / "geo"-centric play (to borrower the term used earlier) as well. The situations in which I'd outright deny a player declaration are almost zero. And of course there's always spin-off / downstream consequences for what they do.
But even if they players declare this action, the framing for the "steal the ship" scene is still largely going to be of my devising---unless I wholesale grant the players the ability to do some of the framing themselves.
If I wasn't willing to give the players some of the fictional creation / framing power, I had to do it myself.
What kind of ship are they trying to steal? Who owns it? What's the owner's relationship to other people in power inside the city? Who's guarding it? How well is it guarded? What's on the ship when they steal it? How easy is it to access the dock? Is the party likely to be pursued afterwards? Who will the pursuers be and how will they be engaging in the pursuit? What happens if they're recognized at another port of call? Who recognizes them? <
ad infinitum>
There's just so much detail that falls out from that single action declaration---"I steal a ship from the harbor."
And so many of the answers to those questions ultimately become "stuff in my (the GM's) notes"---stuff that the players are going to want to have knowledge of. Because players don't like to do stuff without understanding the risks, understanding their potential level of success, whether stealing the ship actually has a net positive or negative outcome on their goals (or fulfills some of the goals while undermining others, etc.). Someone at the game table has to ultimately generate these kinds of details for the fiction / framing around the proposed scene.
If it's all the GM's call to determine these details, then a significant portion of the players' actions are then going to be just what
@pemerton described, which is, they're now playing to
find out what's in those notes so they know how and when to actually initiate their "steal a ship from the harbor" action declaration.
I'm always willing to pivot and let players pursue new courses of action. But there's a ton of ancillary "note generation" that then goes into it, and those notes have a significant impact on play.
*Edit---this is the kind of thing I was talking about in my earlier post about how making the players the focus of the action "required a significant amount of negotiation / 'Mother-may-I?' or outright 'handwavium' to make them the focus of play." All of this stuff around framing the scene for stealing the boat is stuff that I basically have to generate---and if I-as-GM am the only one creating those notes, then I either have to outright tell those players what I've noted, or the focus of play now becomes figuring out what I put into those notes.
Even with the best of intentions, the ability of the players to successfully carry out their course of action is all based on a GM judgement call of
what did I put into my notes?
This is what I was trying to communicate to
@Emerikol previously, which is that no matter how detailed your initial "prefabrication" is, these types of details around individual scene frames (like "We steal a boat") are not pre-existent when the action declaration is presented, or even if they are,
they're still "notes" that the players have to now retrieve from the GM before they can realistically make the "I steal a boat" action declaration.