So, one thing I think is cool is how Undertale’s mechanics make it work in a way that’s not immediately obvious. I’m talking specifically about the combat mechanics: the timing minigame of attacks and the bullet hell minigame of defense, the latter in particular. At first it seems like a random mechanical gimmick, like many RPGs have. But in addition to providing cool, nonliteral opportunities for flavor, these minigames more importantly provide support for multiple paths.
Game mechanics can involve player skill, character skill, or both. A game like Mario is basically player skill. It has some powerups, but basically the game gets gradually more difficult and you improve as a player to compensate. Many RPGs focus on character skill: there are player skill elements to greater or lesser extent in terms of choosing what to use when, but the effectiveness of attacks and defense is determined by character stats (plus randomness) and the way to react to improved challenge is by leveling up.
Undertale makes either path possible. You can level up your character by killing, coping with more difficult challenges by dealing more damage and taking less damage, or you can choose not to kill, requiring you to improve your bullet-dodging skills as player in order to survive encounters. The choices you make aren't just multiple-choice morality, they actually dramatically change what type of game you're playing. (I'm sure they factor into various Chara/Frisk/player separation analyses as well.) If Undertale used a normal RPG combat mechanic, the pacifist route just wouldn't work: at some point you'd reach an encounter that was mathematically impossible, or at least vanishingly improbable, to beat.
It reminds me some of LARP mechanic choices, most obviously combat. At one end, what Six Levels of Substitution might call Adaptation, you have player-skill based simulated combat: boffers, nerf/disc/dart guns, spell packets. Games with those mechanics often have character stats that theoretically make some characters better at combat than others, but in practice experienced players who are physically fit and/or well-practiced at the particular mechanic often dominate because, like in Undertale, stats matter less if one party can dodge well and one party can't. On the other end you have maybe turn-based stats-only or stats-and-dice combat mechanics, where the chances of a character winning a particular fight are determined basically entirely by character stats and items, like the traditional RPG mechanic. (Waylay is basically all player-skill, just misdirection and distraction rather than physical abilities. Darkwater combat is somewhere in-between: it has a large character skill component, but in many cases reaction time and coordination with allies determine the winner.)
In LARPs, it ends up being similar to the Gamist/Simulationist divide. If you view the LARP as a challenge that you're trying to win, using all your player skills to do the best you can is obvious. If you view it as simulating a particular world/situation, then it doesn't make sense that the frail little orphan can beat the world's greatest swordsperson in a fight. (It makes the swordsperson's skill an informed attribute/have weak Mechanical Support.) Things to keep in mind if you're trying to reduce how much players are likely to ruin other players' fun. Either way, though, most LARPs aren't trying to play up the player/character divide the way Undertale does, so it's good to think about if you're setting your game up for dissonance. (If you're trying to do a player-skill-based game/mechanic, not having character skill be strongly established can help.)
(It's a bit interesting bringing this Simulationist perspective back to video games. Mario shouldn't be able to beat Bowser, he only does it because of Bowser's poor choice of fighting arena. Whether LV 1 Frisk should be able to defeat Asgore depends on what assumptions you make about the characters. But in general, the ability of PCs but not NPCs to level up makes numerical combat in video game RPGs a lot different from numerical combat in LARPs.)
It all goes back to thinking about what you're trying to accomplish with a given mechanic. Undertale obviously put a lot of thought into its mechanics and how they interact with what the game's trying to do. Whether you're importing a mechanic or trying something new, think about how it will interact with other elements of your game, different player types, and your overall goals, and act with intention.