Proof fields not used.

`proof.isSane()` checks for nil values for the type `deadend`. Is there even a need for this? Looking through the code it appears there's no possible way for `key,value,hash` to ever be set for `deadend`. Is it simply good enough to check for the type without testing the unused fields? Or are these fields placeholders for future reference?

Locate proof.isSane() and the deadend proof representation, then trace whether key, value, and hash can ever be populated for that type. Review how deadend proofs are constructed and validated. The work is complete when the intended handling of these unused fields is decided and the validation behavior consistently reflects that decision.

Written by the indexing model from the issue text.