Security Through Clearing: The Timelock Model

Transcript

[music] [music] >> Thank you for that introduction. Um, today I want to propose something a little bit unpopular, um, which is to to take some ideas from traditional finance and apply them to crypto to uh make our ecosystem safer. So, let's start here with a credit card. Um [snorts] when you swipe your credit card, there's a series of intermediaries uh in the middle and the settlement and the authorization are multiple days apart.

So, when you swipe the card, the transaction doesn't settle, um, and there's there's quite a large uh window as a consumer where you can go and claw back and cancel a transaction. Um, and the same thing is true for ACH. Um, wires you have like a a 30-minute period where you can cancel, but these are mostly uh instant. And in crypto, basically all we have are wire transfers.

We only have transactions that settle instantly with no chargebacks. Um, we have all these different hacks and we also have, you know, add Binance to this list, too. And in all of these cases, the transaction and the signature that the user signed, it was valid. Um, they just had no way to know or understand really where the money was going unless they were uh super technical. And then there's no uh takebacks. So, what I propose here is basically adding um multiple stages in a transaction life cycle to protect the end user.

And I think that one of the really hard things about this kind of model in security is that no one ever thinks to themselves like I'm going to get hacked. Everyone always thinks hacking and you know, getting scammed like that's something that happens to someone else over there. Um And so really it's a it's a large behavioral um thing to overcome and if you have some sort of technical framework where transactions are not instantly um settled, then it allows you to in the delay window for the transaction to actually cancel it and get your money back to safety uh in case you signed something malicious.

Um So yes, I propose that we copy this part of the banks where we have a period of time where uh actions they will wait um and in this period you're able to cancel and you also have maybe a third party that can cancel on your behalf and you wouldn't let them uh cancel all the time, you would let them cancel maybe only one or two times so they can't permanently lock you out of your funds. And uh you know, like the previous uh presenter Dev proposed a policy engine, you could add um calls that would be able to bypass this time lock.

And the same thing should apply also to account recovery. Um anytime you can do an instant action that there's an opportunity there for an attacker to strike. And so the more delays and waits and periods that you can put between users uh signing and something actually happening, the more likelihood you have of um people not being compromised. Okay. Okay, yes. So a time lock it introduces some new challenges um when you're trying to perform swaps, um people can front run you and anything that's non-routine is going to

get a lot slower. If you don't have someone that's watching these time lock transactions, means you really don't have any protection and if every signer approves something that's malicious and you don't have that watching, um the money is still going to get stolen. So, a challenge that I would give to the audience here is to to try a time lock. There are a few different time lock wallets on the market. No such safe has one, I believe, called Fiducia and we've built one called Cleedy. And try it out and see how it works because there's a lot more friction using this than normal wallets.

Thank you. And you can follow follow me on Twitter. >> Okay, thank you very much. So, basically you propose like small delay button in the transactions, right? >> Yes. >> Okay. >> Yes, indeed. >> So, let's see if we have any questions and we have some questions. So, let's start with the first one. How does the clearing period work when we have time sensitive transactions? Would it not be counterproductive? >> Really good question. So, having this clearing period when you need to do something immediately is going to be annoying and so you'd want to have a policy engine that allows you to bypass the time lock for actions that you know

are safe. But if you have an action that's outside of that, um outside of the white listed actions, you're going to have to wait, which is you know, not ideal. It's kind of a trade-off. >> So, we can do one more question and this is this is interesting one. How would this approach prevent a hack like the Bybit one? >> So, this is a really really good question and the answer is um you would for Bybit, you have this malicious malicious transaction but instead of it being immediately executed on chain, it would now go to a waiting

period where anyone could observe it. And if you had sufficient monitoring during that time lock period, the monitor would flag and say, "Hey, this transaction is going to upgrade the gnosis safe to an unapproved and unauthorized contract and you need to do something and you'd be able to go cancel that transaction." >> Okay, next one. Um, doesn't having a clearing period expose a transaction to a completely new type of attack? >> Um, you can get front ran if you're if you're doing um, swaps on a on a public time lock.

Um, yeah, that that that can definitely happen. So, you'd have to move the swaps to a different wallet. >> All right, so our time is uh, almost up. So, thank you very much for the talk. It was very inspiring and please >> [music]