I am not endorsing this idea but I am bringing attention to it because it sounds like a fun way to fight Citrea and Stamps:

Modify bitcoin core to enforce a mempool policy that rejects transactions that put money in addresses without including a proof that each address can be spent from. The proof doesn't go in the blockchain, it's just relayed with the tx to prove its outputs are spendable.

This would require wallets to create the proof when generating an address and share the proof with the sender. It would also require multisig wallets to use some sort of coordination mechanism to produce the proof. But it requires no consensus changes.

User flow: Alice wants to send money to Bob. Alice asks Bob for an address. Bob's wallet not only generates the address, but also signs the message "Proof of Spendable" with his private key and shares the signature with Alice, in the same qr code as his address. Alice's wallet scans the qr code and constructions a tx paying Bob's address, and when it sends the transaction to the mempool, it shares Alice's signature along with it.

Nodes only propagate the transaction through the network after verifying the signature so that they know Alice can actually spend the money in that address, meaning it's not just a data carrier. Stamps and Citrea, RIP.