No game is "about topological sorting", because that, in itself, is not fun. Fetch quests in general (or chains of deals), are hated by most players. They're tedious, repetitive and arbitrary, with barely any connection to the central storyline (you'll notice the two games I mentioned earlier were both quite poorly received). If you google "fetch quest", the first thing that comes up is "Fetch quests are often considered the lowest form of RPG sidequest", and that's not wrong.
However, plenty of games make use of topological sorting, and if you already have some other kind of game that is fun, then these procedurally generated fetch quests are a great way to increase replayability or add extra content. It should absolutely not be the main concept though. Procedural generation is a defining characteristic of roguelikes too - but without monsters to fight, there would be no game there.
At the moment, there is no challenge. The only way a player could possibly lose the game, is if they get bored and quit. The easiest and most common way to add a challenge, would be to add combat. If you make it realtime, you have a procedurally generated Zelda-like. If you make it turn-based, you have a non-linear Roguelike (this would be the simplest). I'd play either of those. The other alternative is to add puzzles.
Ideally, though, try to avoid simple fetch quests to begin with, and make it so the items you find are tools that the player can actually use in some way, and not simply exchange for some other random item.