Greater resistance against quantum attacks that directly target the blockchain’s infrastructure.
Your changes don't affect quantum resistance in any way.
We generally discuss quantum attacks on blockchain systems with a focus on the safety of signatures, but rarely we talk about the threat of something like quantum-miners, I mean a quantum processor that gets used as a miner device. with this point of view, when we NORMALLY follow our daily mining process and sometimes also find a negative-hash and write it in the block, an attacker needs to ENFORCE his processing-equipment to find a match to satisfy the network's difficulty in both regular and negative ways for a culprit reorg to pass that special block. that's how the idea could help better resistant against quantum attacks.
but including them in Bitcoin's block requires a protocol change, which is unlikely
We may not need to change any thing about block header.
Just imagine I have found a new block with a valid negative-hash attached to it (not included yet), so I could broadcast it to other nodes as always we did before. for the next block, while we have the previous block's hash in the new block header, miners could also find a proper place in COINBASE transaction for the previous negative-hash value (if exists) and chain it to the new block header by its regular merkle-root. This is how such special negative-block could simply get fit in Bitcoin blockchain with less changes in major aspects of the current working Protocol.
P.S.: Just as a primary incentive, this also could help we access our coinbase rewards / fees with less confirmation.