Titan Extreme

Then do the changes and we're set.
Clear doesn't affect dodge value: if it's physical attack, then clear will ignore it, no matter what.
Landslide removes the party? perfect. Let's make it happen, it's starting to become tasteless my victory, without following the theme of your battle.
If you won't ban Black Belt, then put the trigger in the attack routine too, even if i may trigger the next phase with a normal attack, i may do it with a counterattack too.

Actually i knew since the beginning when and where they attack(when landslide hits a single ally and not 2, like in a 2 vs Titan isn't anymore a great menace, afterall), maybe not so much about the counter, due to the discovery of the counter attack(i can study a bit the script, anyway).

Do what you want, but please, fix as soon as possible the bug, i won't try the battle anymore, until you fix it...
I cannot stand a battle crush after i'm at the end of the fight

Playing passive won't make it trigger, because you don't have to move, but just dare to act while Titan makes his turn, and the bug will hunt you.

Actually it's triggered by special, doing almost anything: swapping current turn with the one of another ally, opening item, dealing with magic list(one among open that window, or closing it) and probably decide the action and so pass to the next ally turn.
Mt. buster does it too, but it's less frequent.
Even landslide seems that may mess sometimes with the bug.

At this point i might say that any of your custom animations could trigger the bug(i'll never test Earthfury anyway ), just i don't get why the special too, but maybe the animation type could belong to the special's animations that gives some problem to the game? test that animation in a normal FF6 rom, in the case.

One time it happened something really weird: I triggered the bug, then, after a while, it suddenly appeared Interceptor's static sprite next to Cyan(maybe), it had a different palette... after few seconds, seems like that triggered death party landslide(if i recall, Interceptor's status check was inside the attack routine too... ).

Does this stuff gives some clue about the bug? I hope it so...