Mostly agree.
How damaging will Hama/Mudo be in relation to Kou/Ei?
Pierce is somewhat inelegant (inefficient) design, making Resist and Null redundant by endgame such that game elements exist to counter each other for no net effect. The bigger issue is that damage negation exists at all and/or enemies are stupid, such that the player can stalemate enemies and enemies can stalemate the player, violating the ideal that the every action should always move the game towards its conclusion of winning or losing the damage race.
The new Gun skills are good, since the original ones were dominant even without Boosts.
Critical Eye has many implications, potentially narrowing the variety of optimal Phys builds to those with the highest damage plus Critical Eye, ignoring innate Critical rate. It also makes superboss ||Lavenza|| phase skips more consistent with Hassou Tobi Critical on Rage.
Ignoring Critical Eye, 0% Critical rate Hassou Tobi and God's Hand, less damage on God's Hand, and more damage on Gigantomachia are valid changes.
The Archangel mini-boss having Critical Eye over Charge implies more turns, which should be good, but could you consistently negate One More by guarding? I don't remember.
The Archangel mini-boss served another purpose: an SP check; assuming you battled all enemies and hit all of their Weaknesses prior, you should have near 0 SP at Archangel, though spamming Gun saves SP. Do you plan to address this? (Not necessarily good or bad.)
Why remove the Breaks and Walls, Baisudi/Mabaisudi, and Last Stand and Adverse Resolve? Because they were inconsequential, or to remove rewards for mistakes? Is design that promotes meaningful decisions possible with these game elements?
Why not remove Almighty Boost?
Slime Mara having Diarahan and an Ice Weakness to force MC to use Ice to rush one Mara is valid.
Kamoshida changes validly incentivize MC as damage dealer.
Shiki-Ouji was a hard SP and damage type check. Does softening him make any sense?