Claude's proof just uses Mathlib's formalization of Riemann's Hypothesis. It's the closest thing you can find to an "agreed" one. It's not like that Claude tried to formalize it from human language from scratch.