I'm not sure these all qualify as "mind games," but I'll list a few. In order of how useful they are from what I can tell:
Top 4 imo:
1. Dash/run into their burst range and wavedash back to react to what they do.
2. If they tech or whiff an attack just out of reach for you to punish, run straight at them like you think you can make it in time (even though you know you can't). Run past them and pivot grab. Since they think you misguessed what you can cover, they will likely shield or spot dodge hoping to counter you (especially since they have a frame disadvantage if they go for attacking). Since they'll wait for you to hit the shield if they shield, and since running past them delays the grab to be after when they would spot dodge, this covers their two most likely options.
3. If they tech on a platform but you are out of reach to tech chase on reaction, then waveland on the platform as their tech comes to an end and go for a grab. They are again at a frame disadvantage for attacking, and spot dodge this time is a worse option for them because moments earlier you were in the air (so they'll fear a lingering hitbox), and because you being in the air makes them fear grab less. Rolling is uncommon because they are on a platform, making rolling worse than normal.
4. Short hop fast fall like you are going to do a spaced aerial (so don't land too close). Instead just land and run up and grab.
Also useful:
-Jump at a high aerial opponent to make him fear a juggle attempt then just fast fall and position yourself under them. This can possibly bait their second jump, which would allow you a better juggle opportunity once you land. This could also cause them to attack early and whiff, which may allow you to punish them when they land and/or by having you jump back up to hit them as their move finishes.
-When at the ledge, repeatedly regrab the ledge and/or fake get up attempts until they commit to something they shouldn't. Use that opportunity to get back on stage and/or punish.
-When coming down like you are going to do a spaced aerial, instead waveland backwards and react to what they do.
A couple things to note: A lot of these mind games focus on punishing smart options from opponents. I would say you can go for these right off before observing your opponent to know that they will choose the "safer" options, but still keep an eye or for things like them attacking despite having a frame disadvantage ect. Sometimes people choose unsafe options that ironically counter some of these tactics. If this happens, notice it and switch away from the tactics this interferes with and focus more on punishing their unsafe habits. Generally speaking, the less committal an action, the better it is for punishing the less safe options. #1 doesn't rely on the opponent's choices much and the tactic is far stronger if the opponent chooses unsafe options often. As for the issue of adapting: your opponent is dumb until proven smart. Once you've seen solid evidence that they are attempting to adapt to a tactic, mix up with less committal things on occasion (like wavedash back) so you can punish the less safe options as well.