If we had Merkle trees in Bitcoin script, we could publish the Merkel root of a mint's spend book on-chain and give users who can prove non-inclusion a way to withdraw unilaterally.
Interesting idea. Merkle roots of mint spend books on-chain would give users cryptographic proof of solvency without revealing individual balances. Reminds me of exchange proof-of-reserves, but for decentralized mints. The challenge is on-chain cost of frequent root updates during high volume. Would need efficient batching or an L2 bridge for practical deployment.