I've been digging into Mostro's P2P trading model. The part that interests me most from the Lightning side is the use of hold invoices.
Instead of depositing BTC into a conventional exchange account and receiving an internal balance, the seller commits sats to the trade through a Lightning hold invoice. The HTLC remains locked while the trade is completed, with settlement/cancellation following the protocol flow.
Obviously this doesn't remove every trust assumption. The coordinator, Lightning implementation and dispute process still matter.
But I think it's an interesting example of Lightning being used for something beyond simply “fast Bitcoin payments.”
I'd be interested in criticism from Lightning operators in particular: where do you see the weakest assumption in this model?
I think that the weakest assumption still is that you aren't trading with a coordinator that is also the arbiter on disputes.
That doesn't mean that the coordinator is a bad actor per se, but coordinators can get hacked, and once a nostr key leaves your custody, you are fully screwed.
Afaik Robosats also have the same non-escrow model like this no ?
My concern on this hodl invoices is that if mobile wallet user left their wallet offline (might be beginner not enabling persistent mode) when the HTLC got cancelled what's going to happen to their collateral tho?