Red-Teaming Crypto Wallet Operations

Transcript

Okay. Uh, so sorry I actually rugpulled my topic a little bit. Um, still around red teaming, but uh, specifically around this idea that I've been brewing around called the crypto transaction killchain. So what's the parallel to red teaming here? Uh, in traditional kind of networks, red teaming, uh, we use this thing called the cyber killchain, right? This is a really helpful uh sort of traditional model and lexicon for everyone to sort of speak the same language to understand you know how are networks compromised um how do we need to defend against them and [clears throat] then ultimately what are the attackers objectives right to be able to map this all out. Um, so this is super helpful for networks, but networks are really just, you know, data, right?

Blockchains are a little different, right? Blockchains are all about value, right? Yes, there's data. Yes, there's different things in there, but it's really, you know, a system of value. So, I see a lot of people have tried to map that cyber killchain to crypto attacks. I don't think it works. So what I'm proposing here is this new uh killchain specific to crypto and would love to get some ideas and feedback from everyone here on on uh all of this. Right? So uh the the whole idea here is that you know in crypto the objective is different right the objective of the attacker is to steal money. It's not you know ransomware. It's not stealing data. Um you know there's lots of ways that they can do this. could be theft, manipulation or exploitation, but it's a

totally different objective than you know a traditional network compromise, right? So your threat model should also be different, right? So this starts from, you know, I kind of think of this as who is the human first, right? What is their intent? Is that a good human? Is that a bad human? Are they encoding that intent or signing that intent with good or bad software? Are they signing that intent with good or bad hardware? right? Uh is this legitimate? Is this backdoored? Um and really the whole sort of transaction flow from the point where you know, hey, it's either a good user that initiates that transaction or a bad user that initiates that transaction.

What are the types of controls that you can put in in that kill chain to mitigate something bad happening? So whenever I read a postmortem uh of any attack, um my thought process is you know how could we have mitigated this, right? Um so you know for uh the buy bit attack right was safe right things went bad right kind of up there in like the transaction approvals right the users got prompts it said hey this is a normal you know transfer that we expect and they blind signed it and then ultimately the signing was compromised right uh the signing interface was compromised so to me this

is just a very helpful way to kind of think of all the areas that you can apply really checks and security throughout this whole process. So it starts with very simple things like all right making sure you know my developers aren't North Korean IT workers at the top. uh it's making sure that you know in my wallets and my kind of uh security right I have strong role-based access control strong private key security maybe multi-party computation um it then goes to you know how do I actually read or write to the blockchain right if I'm reading from a uh uh tampered with RPC or a you know malicious DAP right uh is that an opportunity to to either detect or

prevent something and then ultimately Right? Once you know at Fireblocks, we really focus on you know mostly the the top part there, right? Because we're a wallet company, right? But you can augment the wallet with all of those onchain controls. So, you know, the gentleman right before me who was talking about, you know, safeguards, that's a great one as well, right? So, once you sign a transaction, that shouldn't mean game over, right? There's still ways to prevent and detect uh and, you know, prevent that from going through. And so that goes all the way through, you know, to really just transaction finality, right? And what does that mean? Is it a permission chain where you can actually control a consensus and like roll something back?

You know, we don't want that. But just think, you know, again, just like a mental model here to kind of think about this uh all the way through uh intent to finality. So I don't think any other frameworks again like would love to get feedback if if anyone thinks something like this already exists but like the ones that I looked at like the blockchain security standards council and the cryptocurrency certification consortium uh they're very like objective based like you must do this you must do this uh it wasn't really like a good you know kind of traditional killchain uh that kind of maps like what do attackers do and where are the opportunities to detect and prevent them.

So yeah, so would love feedback uh from everyone here, right? Uh is somebody else working on this? You know, can we collaborate? Like I'm not married to any of the naming or any of this. So if you have ideas, I eventually want to open source it uh and then uh you know, start expanding it to something like a MITER attack framework. That's all. >> Awesome. Awesome.