Let’s try to understand
Schema hooks get interesting the moment a protocol stops just recording claims and starts shaping what is allowed to happen. If Sign lets custom logic sit inside attestation flows, then where does protocol responsibility end and application responsibility begin? If a hook rejects, validates, charges, or triggers something, is that still neutral infrastructure or already business logic wearing protocol clothing? And if every schema can behave a little differently, does that make the system more composable or just harder to reason about under audit? That is the part worth watching. Power is useful, but blurred boundaries usually come with a cost.