DSS 2026 — Security is a leadership problem Josselin Feist Source video: https://www.youtube.com/watch?v=x4-TYxCvwvE Automatic captions — not human-reviewed. Complete spoken-audio coverage is unverified. Source language tag: en (English). Source export SHA256: 2e8205e1daa14ce3ddfff68c76f91ba5ed000f06dd55fd5f576c0e1bcdf85d26 [0:01] Okay. Hi everyone. Um, welcome to this [0:03] presentation, security is a leadership [0:05] problem. So, first thing first, who am [0:07] I? My name is Jocelyn Feist. I'm a [0:09] security researcher and I've been in the [0:11] blockchain space for almost nine years [0:12] now. [0:14] Through the years, I had the chance to [0:15] work with many protocol, DeFi, layer [0:17] one, layer two. Um, and I have worked [0:20] with protocol with various stage of [0:22] development, various maturity. [0:24] This presentation is a personal take [0:26] based on some of the observation I have [0:27] made through the years. [0:29] Oh. [0:30] Sorry. [0:32] So, when I started my career as a [0:34] security researcher, my main focus was [0:36] to find bugs. My kind of thinking was [0:39] that if I can find all the bug in the [0:40] code base, my job will be basically [0:42] done. Security is a technical problem, [0:44] so I need to find technical solution. [0:46] Over the years, I've realized that many [0:48] bugs stem from incorrect design [0:51] decision, architecture decision, or [0:54] shortfall in the testing practices. So, [0:56] I ended up focusing a bit more into [0:58] helping protocol and developer to think [1:00] about their design and the long-term [1:02] security. [1:03] Uh, that helped, but over time, I've [1:06] realized that this is not enough. [1:08] Even if you find improvement you have to [1:10] make in the design, even if you find [1:12] shortcoming in the testing, someone has [1:14] to make a decision. Someone has to [1:16] decide, are we going to allocate [1:17] resource? Are we going to change the [1:19] product? Are we going to change the [1:20] timeline? And this made me realize that [1:23] security is not just a technical [1:25] problem. And this is sometimes a bit [1:27] difficult because as engineer, we focus [1:29] on technical problem. But at the end of [1:31] the day, uh, a lot of the decision are [1:33] going to be made by the leader of the [1:34] protocol, and as a result, security is [1:36] not just a technical problem, it's also [1:38] a leadership problem. [1:41] We knew it was a weak point, but we [1:44] thought it was okay. That quote stuck [1:46] with me. I have heard it many time from [1:49] different team with little variation [1:51] after an incident. [1:53] When you were looking back, when when [1:54] the team was looking back why an [1:56] incident happened, they realized that it [1:58] was not something completely obscure [2:00] that happened. It was something that [2:01] they knew about, but thought it will be [2:03] okay. That could be a low severity issue [2:06] that was acknowledged and never [2:07] revisited. That could be a situation [2:10] where the smart contract were actually [2:11] robust and and were in a good point, but [2:13] the team never considered operational [2:15] security and as a result, that's why the [2:17] attacker [2:19] led to exploit. [2:21] We have in the team we we have in the [2:23] industry really good and strong [2:25] technical team. They care about [2:26] security, they spend a lot of time, but [2:28] sometime they still fail because at the [2:30] end of the day someone has to make a [2:32] decision, someone someone has to decide [2:34] are we going to go live or are we going [2:36] to spend more time looking for issue. [2:39] And this is a leadership decision. Now I [2:41] know you might have heard similar take [2:43] in the past and I know leadership can [2:45] sound like a bit of a disconnected [2:48] topic, but I so know firsthand that many [2:51] protocol leader in the space do not [2:53] consider security as their [2:54] responsibility. [2:56] As a result and as we are going to to [2:57] see through this talk, most of the time [3:00] issue are never revisited, risk are [3:02] never revisited, exposure can scale [3:04] faster than the confidence in the [3:06] system. There is no security ownership [3:08] within the protocol team and more than [3:11] often perception of security becomes [3:13] more important than the security itself. [3:17] So who here has deployed a protocol [3:20] within the past let's say 12 months and [3:23] could point me to the list of non-issue [3:25] on the list of risk accepted and would [3:28] be able to tell me when they have been [3:30] revisited. [3:31] Probably only okay, only a few people, [3:34] right? [3:36] In reality, you are going to make [3:38] trade-off, you have to make decision. [3:41] And you know, when you need to go live, [3:42] you're going to accept some of the risk. [3:45] But in reality, once you accept a risk [3:47] even temporarily, most of the time they [3:49] become permanent. [3:52] And the problem is not necessarily that [3:54] you make a trade-off, the problem is [3:55] that you never revisit them. However, [3:58] your protocol will evolve. You will have [4:00] new integration, you will have new [4:01] asset, you will have new components. And [4:04] if you don' spend more time trying to [4:06] think about how to re-evaluate and [4:08] re-assess the risk, something that was a [4:11] minor risk can quietly become a [4:12] permanent risk within your system. [4:16] So this is where as a leader you need to [4:20] periodically re-evaluate and and your [4:22] risk and allocate resources for that. [4:24] This is a boring task. No one likes to [4:26] do that. And if you do it, you know, you [4:29] won't be able to tell your investor that [4:30] you have made progress. You won't be [4:32] able to generate revenue. You won't be [4:33] able to kind of like, you know, show [4:35] your user that you have like a new [4:36] report to publish. And most of the time [4:39] when you are going to re-evaluate your [4:41] non-risk, no decision will be made [4:43] because the trade-off that you have made [4:44] might be still true today. [4:47] So this is a bit uncomfortable, but it [4:49] is why this is a leadership decision. [4:52] Next time you deploy a protocol, you can [4:54] tell your engineer, we are going to [4:56] deploy next week, but in six months from [4:58] now [4:59] we are going to re-evaluate your risk. [5:01] This implicitly will tell your [5:03] engineering team that they need to [5:05] document and own the risk. [5:07] And this is how as a leader you can [5:08] create a culture in within your company [5:11] where the management of risk is [5:12] important. [5:17] I frequently hear protocol leader tell [5:20] me that they don't have the budget for [5:22] security. [5:23] Now security is not just about paying [5:25] for audit. It's also about controlling [5:27] your exposure while your confidence in [5:30] the system is limited. [5:32] Most of the protocol team won't have the [5:34] budget to do multiple round of audit, [5:36] formal verification, fuzzing, and like [5:38] all the different techniques. But they [5:40] can decide to not expose immature system [5:43] to unlimited loss. [5:46] We have seen multiple time in the space [5:48] small experiments, small idea that [5:50] overnight become a high value target. [5:52] And this is a problem if your security [5:54] budget and your security confidence in [5:56] the system does not allow you to [5:58] properly assess your system. You do not [6:00] want to be in a situation where your [6:02] exposure scale faster than your [6:03] confidence in the system. [6:05] The good news is that as the protocol [6:08] leader, you have a lot of lever that you [6:10] might never know you have. [6:12] For example, when you are going to [6:13] create your first version of the [6:14] protocol, you can you can think about [6:16] reducing the scope of what you are [6:18] building. This means we're moving some [6:20] integration, we're moving complex um [6:23] complex part of the protocol if you [6:25] don't need them for V1. [6:27] You can directly limit exposure, you can [6:29] control your exposure by having TVL cap, [6:31] by having rate limits. [6:33] You can have centralization point that [6:35] are will going to help you to control [6:37] that, for example, pausing mechanism and [6:39] this type of things. Now, this decision [6:41] and this trade-off are a bit [6:43] uncomfortable because they're going to [6:44] compete with competitiveness, growth, [6:47] revenue generation, but this is why they [6:49] are leadership decision. [6:51] If you don't do it, no one will do it. [6:53] And this is where great leader think [6:56] ahead and think about what they're going [6:58] to build and how their security budget [7:00] will allow them to build what they want [7:02] to build. [7:04] Now, this can become difficult to [7:06] evaluate if you don't have any security [7:09] expertise and any security ownership [7:11] within your company. [7:13] If you look at traditional financial [7:15] institute, they have a lot of security [7:17] expertise. They have a CISO, they have [7:19] an internal structure, they have [7:20] security DevOps, they have security [7:22] engineer. [7:23] If you ask most of the DeFi team today [7:26] how they handle their security, they are [7:28] going to tell you we are done audit. [7:31] Most of the industry has externalized [7:34] security expertise to third-party [7:36] provider. [7:37] That model made sense when the industry [7:39] was scaling and expertise was scarce, [7:42] but it is not a good model today. [7:44] Protocol team, blockchain team, and DeFi [7:46] team needs internal security and they [7:48] need security ownership. [7:50] Now, what I have see sometime happen is [7:53] that one engineer on the team become the [7:56] owner of security, but that engineer has [7:59] no lever. They cannot make decision, [8:01] they might identify risk, they might [8:02] identify concern, but they have no [8:05] ability to change any direction in the [8:07] project timeline and in the project [8:09] features. [8:10] Security, like ownership become mostly [8:12] cosmetic. It exists on paper, but [8:14] doesn't exist in reality. [8:16] This is why you need security ownership [8:19] and you need this authority, this [8:21] ownership to come with authority. [8:23] How to do that? A good model I've seen [8:26] work for early stage company is for one [8:28] of the founder to own security. [8:31] They will decide how resources are [8:32] allocated, how the product timeline is [8:35] going to be aggressive. If a refactoring [8:37] is needed, they will be able to say we [8:38] are not going to go live until A, B and [8:41] C is fixed. [8:43] Now as you scale, you might be able to [8:45] delegate that ownership and that [8:46] authority to someone else and to create [8:48] a structure for that. But many of the [8:51] team that have good security posture [8:53] today started with that model, sometimes [8:56] even without without realizing it. [8:59] So now you're in a situation where you [9:00] know your risk and you re-evaluate them. [9:02] Your exposure scale at a at a pace where [9:05] you can control it and you have security [9:07] ownership with clear authority. There is [9:09] another aspect where I've seen sometime [9:12] leader failing without even realizing it [9:14] is how they communicate about security. [9:18] Who here have been in a situation where [9:21] you see a security report, there is a [9:23] high security finding [9:25] and the leader of the company or the the [9:28] lead of like the protocol team [9:30] is pushing back to reduce the severity. [9:33] So you report high severity and now you [9:35] enter kind of like a back and forth with [9:36] the developer team to reduce the [9:37] severity to medium or something like [9:39] that. [9:40] This creates [9:42] a focus on the perception of security [9:45] and not the security. It's it tells the [9:47] engineer that the optics of how the [9:49] report will looks like is more is more [9:51] important that how the security issue [9:54] might have impacted the team. [9:56] I have been in a situation where I was [9:57] reporting high severity to teams and I [10:00] could see that the engineer was stressed [10:02] because he was afraid of the consequence [10:04] of having a high severity in this [10:06] report. [10:07] And this create again a situation and an [10:10] environment where the perception of [10:12] security is more important than the [10:13] security itself. [10:15] No, mature team behave really [10:17] differently. The leaders want the [10:19] assumption to be challenged. Where you [10:21] do security review and an audit with a [10:23] mature team, the engineer will seek [10:25] feedback. They will not be defensive [10:27] about their code. They will want to you [10:29] to challenge the different assumptions [10:31] that have been made within the code. [10:33] When an engineer [10:36] when an engineer see a concern, they are [10:38] going to raise it internally. I am not [10:40] confident about this component. I think [10:42] we should spend more time. [10:44] If a refactoring is judged necessary, it [10:47] is going to happen. If it needs to delay [10:49] a feature or a product launch, it is [10:51] going to happen. This means that [10:53] security is not just something that is [10:55] for optics. It does have impact in the [10:58] decision. [11:00] I have been this is in a weird situation [11:03] where I was reporting what I thought was [11:05] a medium severity to a protocol team [11:08] and the engineer started to argue with [11:09] me about the severity, not to decrease [11:11] it but to increase it. So I'm like this [11:14] is medium and the engineer is like no [11:15] no, this is high. [11:17] This can only happen in an environment [11:19] where the engineer feels safe to raise [11:22] concern and uncertainty instead of [11:24] trying to minimize it. [11:26] This is why as a leader, the way you [11:28] communicate with the team is really [11:29] important. Whatever you say is going to [11:32] have a large impact within your [11:33] engineering team and you want your tone [11:36] I want security to favor [11:40] raising uncertainty, raising risk [11:42] instead of trying to minimize it. So [11:44] next time you receive a security report, [11:46] if you see a high severity, do not focus [11:49] on how we can make it look better for [11:51] the end user. Focus on [11:53] what can we learn from this from this [11:55] finding. For example, ask your team [11:58] what do you need? How can I help you? Do [12:01] you need more time testing? Do you need [12:02] more time thinking of design? What are [12:04] the things we can change within our [12:05] process for this to not happen instead [12:08] of forcing focusing on security? Again, [12:10] the optics might have some impact in [12:13] your product, but in the long term, the [12:16] if you focus more on the perception of [12:17] security rather than security, you will [12:19] not build an environment where concerns [12:21] are properly addressed. [12:25] So, it's 2026, we have to talk about AI. [12:29] Some people believe AI will solve cyber [12:31] security and AI will find all the bugs. [12:34] This is not what we are seeing today. AI [12:37] does decrease the cost of finding bugs, [12:39] but everything we just discuss, risk [12:41] management, perception of security, [12:44] exposure scaling, um and security [12:47] ownership, all of these still matter in [12:49] the world where AI help us to find bugs. [12:53] For critical code, I see many leader [12:57] make the mistake of focusing on the AI, [13:00] on how it can scale and make uh engineer [13:03] write faster code. The question become [13:06] how much can we ship, how fast can we [13:08] ship. [13:09] Uh and this is a problem because if your [13:12] output and the capacity and if the [13:14] number of line of code they are [13:15] producing out weight your capability to [13:18] manage risk and understand how your [13:20] protocol growth and like the different [13:22] limitation, you are more likely to have [13:24] incidents. [13:27] Now, as a leader, you can take a [13:29] different strategy. You can take a [13:30] different optic for critical code. [13:32] Instead of focusing on how AI will make [13:35] you faster, focus on how can you use [13:37] this extra capacity to improve your [13:39] security. [13:40] This change how you measure the impact [13:42] of AI within the company. You can know [13:45] focus on okay, are we seeing a decrease [13:48] of security findings during review? Are [13:51] we using new technique like formal [13:52] verification or fuzzing? Do we have more [13:55] control over how we manage risk? [13:58] Thinking of security [14:00] as a first priority in how you integrate [14:01] AI might uh might leads to increase [14:04] speed and that might become a byproduct [14:06] of this focus. [14:08] But for critical code, you should focus [14:10] first on how to improve your security [14:12] before how you can be faster with AI. [14:15] And this is again a leadership decision [14:17] that will change how in general going to [14:19] think and to to shape the code base that [14:21] you are producing. [14:23] So risk management, exposure scaling, [14:27] security ownership, perception of [14:29] security versus security itself. All of [14:32] that, you know, comes down to how the [14:35] leader think about security and do they [14:38] consider security as something that [14:39] shape decision or do they consider [14:41] security as something that is a [14:43] verification check at the end of the [14:45] protocol development. [14:47] And this is difficult sometimes for [14:48] founder because they are not trained to [14:50] reason this way. [14:51] A good um [14:54] a good parallel you can think of is [14:55] profitability. [14:57] As a founder, [14:58] you will not think profitability is [15:00] something you check at the end. Uh it's [15:02] not even if you delegate the execution, [15:04] it is still your responsibility. [15:06] Profitability is going to shape how you [15:08] are building your software, how you how [15:10] aggressive you are, which feature you [15:13] have, which limitation you are okay to [15:14] accept or not. You can think of security [15:17] in the same way. [15:19] Now to build that mindset, [15:21] there are a couple of simple questions [15:22] you can ask yourself. The first one, if [15:25] our TVL grows from 1 million to 10 [15:28] million, what additional security [15:30] measure do we need? [15:31] This means treating TVL growth not just [15:34] as a success but as a risk vector. [15:38] With a security budget, what level of [15:41] complexity and what level of security [15:43] can we responsibly support? This means [15:46] not trying to retrofit your security [15:48] budget to what you have built but you [15:50] build something that match your security [15:52] budget. [15:54] The last two questions are related to [15:56] risk management. So after an audit, what [15:59] what stay possible and uh what [16:02] assumption do we rely on and when do [16:04] they stop to be true? So this is about [16:06] knowing your risk and being having a [16:07] process to reevaluate them over time. [16:13] A lot of what we discussed today were [16:15] around leader um but I would argue that [16:18] anyone within their career can benefit [16:20] from thinking of security as um [16:24] something which has an impact within [16:25] leadership. As an engineer, next time [16:28] you look for a position, choose an [16:30] environment where security matter, where [16:32] the security leader, uh where the leader [16:34] of the of the of the company you want to [16:36] join consider security as something [16:38] important. Ask question, you know, look [16:40] for clues. If you care about security [16:43] and the leader of the company you work [16:44] with don't care about security, you will [16:47] be frustrated. So take this as a [16:49] consideration. [16:50] For investor, here you have an important [16:53] play. [16:54] You can influence your portfolio company [16:56] to think beyond we had an audit. Ask [16:58] them question about the security [16:59] ownership. Ask them question about how [17:01] they manage risk. Um, you have a lot of [17:04] power into how how to change their [17:05] mindset. [17:07] Finally, as a leader, make sure the [17:09] security concern are going to allow you [17:11] to change scope, timeline, and exposure. [17:15] Remember the quote. [17:16] We knew it was a weak point, but we [17:18] thought it was okay. You do not want to [17:21] be the next one to say that. Audit will [17:23] find bug, but your leadership will [17:25] determine if the condition that created [17:28] the issue will improve or not. [17:30] Thank you for your attention. [17:32] [applause] [17:39] Any question? [17:41] >> Hey, I'm Fazeen building [17:43] My question is what acceptance bar [17:46] should leadership accept for a security [17:48] fix before it ships? [17:49] For example, today is just that proof of [17:52] concept fails now. Is that enough? [17:55] >> Um, so can you repeat the question? [17:59] >> What acceptance bar should leadership [18:01] accepts for a security fix before it [18:03] ships? [18:04] Today today is just that proof of [18:06] concept fails now. [18:08] >> Okay, is the question I want um if you [18:11] receive a finding you want to have a [18:13] proof of concepts uh [18:16] uh okay I'm not sure to understand the [18:20] question. Do you mean acceptance having [18:22] you want to accept the how do you know [18:24] you can accept the fees can ship [18:25] basically? Okay, okay. So the question [18:27] is when you receive a fix [18:29] how confident can you be with respect to [18:31] the fix, right? [18:33] Okay. [18:34] That depend on the complexity of the [18:36] finding. If you have a finding and the [18:38] fix is pretty straightforward, usually [18:39] you can have a pretty good confidence [18:41] that the fix will fix it. Now if the [18:43] issue stem from something a bit more [18:45] complex, you might need to ask your [18:47] engineer [18:49] do you think our issue might stem from [18:51] that? For example, if the fix is just [18:53] like [18:55] like an access control missing, that's [18:57] pretty easy to evaluate. If the fix [18:59] require a change in the winding [19:00] decision, that might require to you to [19:03] allocate a bit more resource. Now as a [19:04] founder, if you're a bit unsure about [19:06] that, I will ask the engineering team [19:08] how they think about it because your [19:09] engineer knows the protocol better than [19:11] anyone else. You can also ask your [19:13] security you know expert that internally [19:15] or externally how they think about it. [19:17] But asking the engineering team is [19:19] probably the first place to to to start [19:21] with. [19:25] >> Hello. [19:26] Great talk. Thank you, Jocelyn. So [19:28] question I have is about incidents and [19:31] incidents response. I feel like it's one [19:33] of the scariest things for leaders and [19:36] founders and whatever [19:38] especially in our crypto companies to [19:40] have an incident, but it feels like it's [19:42] inevitable nowadays. And so how do you [19:45] how do you make it part of I guess [19:48] regular day-to-day in business to handle [19:50] an incident, have an incident response [19:52] and all that sort of stuff? [19:54] >> Okay, that's a really good question. [19:56] So yeah, you have to prepare for [19:58] incident because it is likely to happen. [19:59] For that you can think of it ahead of [20:01] time. For example, how you monitor your [20:03] code base not as the code base is [20:06] written, how we are going to monitor it, [20:07] but we are going to build like this new [20:09] component from the design phase. How can [20:12] we integrate monitoring there? What do [20:13] we need to monitor? [20:15] You can test your incident response [20:16] there like war game and and think like [20:18] that. The SEAL which is a Security [20:20] Alliance has a framework website which [20:23] has a lot of different kind of like [20:25] advice on how to do it. So I think not [20:28] considering incident as [20:30] a last step within your development but [20:33] something you integrate from the get-go [20:35] will help you and testing it is also [20:37] really important because you don't want [20:38] to be in a situation where [20:40] you think you have good monitoring but [20:42] actually you don't. Testing monitoring [20:44] is also really difficult because [20:47] depending on the setup you have testing [20:49] like the tooling you are using might be [20:50] tricky so you want to get that ahead of [20:52] time. There are a couple of companies [20:54] that focus on that that can also help [20:55] you you can you know reach out to them. [21:00] >> Hello Jocelyn great great talk. So this [21:03] question is more from the leadership [21:04] side like towards the leadership. So as [21:07] we know like the unfortunate reality is [21:09] many of the protocols they still follow [21:11] the regulations which are already out [21:12] there. However the the velocity of the [21:15] incidence that we are seeing that is [21:17] overriding the regulations that are [21:18] already put into place. So while you are [21:21] mapping out let's say some incident that [21:22] okay a new attack pattern has appeared [21:25] you don't have the regulations for it [21:26] and people are like that okay because we [21:28] don't have the regulations for it maybe [21:29] we not we we are not required to plan [21:31] for it but you will still have like [21:33] incurred losses. [21:35] So how as we as a community ensure that [21:37] the regulations are also you know they [21:40] are like they have a velocity as well [21:42] which is matching up the incidents. I [21:43] think that is really important. [21:45] >> Okay [21:46] that's a good question and a difficult [21:47] one. [21:49] I don't think we should wait for [21:52] regulation to tell us what to do. [21:54] Regulation might help us to have some [21:56] baseline but it is not going to be [21:59] enough and you know we have seen that in [22:02] other industry when when we start having [22:04] like some certification and think like [22:06] that [22:07] it can backfire because usually baseline [22:11] from regulation are not that high and [22:14] they create kind of a floor for all the [22:16] company to follow but that doesn't push [22:18] company to go beyond that and that's why [22:21] building a security mindset for a leader [22:24] is important because if the leader think [22:25] of okay, there is this framework from [22:28] like the regulator, we follow it, we are [22:30] good, that is not going to be enough and [22:32] that's not going to be sufficient. So [22:34] they might help a bit, but I would not [22:36] focus on them. [22:40] >> Um whenever you're dealing with consumer [22:42] protection, um [22:45] uh for example, like you might care a [22:47] lot about your security and making sure [22:48] that your systems are not hacked, but [22:50] consumers can lose credentials, they can [22:52] be tricked into signing into [22:54] applications you have no control over. [22:55] How should should security leadership [22:58] think about protecting consumers even if [23:00] it goes kind of beyond the scope of the [23:02] organization and how do they how do [23:05] should like security leaders try to get [23:06] other peoples in the organization to [23:08] care about it even if from their [23:10] perspective it's going out of their way [23:12] to spend resources on things that are [23:14] maybe not their immediate problem? [23:16] >> Okay. Um [23:18] so like the question is about how how [23:21] leader can think of [23:22] security for the end user which lie into [23:25] like the limit of the protocol itself, [23:28] right? So if it's if it's like on your [23:30] off-chain component like your website [23:32] and think that are connected to your [23:33] website, I would argue it's still within [23:36] your responsibility, right? Now the [23:38] issue is where you have like integration [23:40] with other protocol and when you have [23:42] like [23:42] the limit is a bit more gray. [23:45] Um this is where coordination among the [23:48] industry is important and discussion [23:50] among like the different important [23:51] player is important because you as a [23:54] leader you will be able to secure [23:57] or try to secure your protocol. You [23:59] won't be able to create and enforce like [24:02] some additional mitigation and safeguard [24:04] outside of your organization, but you [24:06] can try to discuss with a different [24:08] partner and with a different provider to [24:10] push in them in that direction. [24:12] Um if you share common investor with [24:15] like other other other company you work [24:18] with, you might talk to your investor [24:19] and try to have like you know a natural [24:21] ground to push some you know mitigation, [24:24] some something. So try for for example [24:26] to find some natural party that can help [24:29] you to push [24:30] you know some additional security [24:31] measure. [24:45] Yeah. [24:46] >> say you're [24:47] a smaller stage company, you know you're [24:49] not an Aave, Morpho, Uniswap [24:52] and you want to communicate. You take [24:54] security very seriously as you, you know [24:57] discussed in this presentation of how [24:59] you need to have an ownership mindset [25:01] and to be able to you know [25:04] take responsibility. We're not just [25:05] going to push it on the auditors. [25:07] How do you then convey that to the [25:09] public if you don't have [25:12] you know the resources to get 10 audits, [25:14] you know to kind of prove that? [25:16] >> Okay, that's a really good question. [25:18] How how can you convey that you have a [25:19] security mindset [25:21] if you cannot pay for you know [25:24] 10 audits, right? [25:26] You you can share some of the things you [25:28] are doing like concrete example of what [25:30] you are you are doing to do that. For [25:32] example, [25:33] maybe you use fuzzing, maybe you use [25:35] like you know formal formal verification [25:37] within your codebase. Maybe you know you [25:39] dedicate some resource time from like [25:41] your engineer to look for vulnerability [25:44] in other codebase or learn from past [25:46] exploit. So you have way to communicate [25:48] concrete example of what you are doing [25:51] to help like the security of your [25:52] protocol and you can can communicate [25:54] that. [25:55] So I will look I will look for concrete [25:57] example that can be translated into [25:59] something public. [26:02] Yeah. [26:06] >> [clears throat] [26:15] >> Thanks. [26:16] With Aave as an example, how do you [26:19] consider [26:21] I guess [26:22] it's [26:23] the wider scope of say hacks on [26:27] cred protocol that end up in Aave [26:30] and and you something that's so far out [26:33] of your scope in many ways. [26:35] >> Yeah, that's a good question, [26:38] because you know, some protocol like [26:39] Aave and other, [26:41] we are impacted when an incident happens [26:44] outside of organization, right? [26:47] And this is scary because we can do as [26:49] much as we can within organization, [26:51] similar to the previous question, but we [26:53] cannot control what's happening outside. [26:55] So it's really about monitoring, it's [26:57] about evaluating who you're interacting [26:59] with. So in the case of Aave, we have [27:01] like, you know, a risk framework, we [27:02] have like a lot of different [27:03] consideration and and additional [27:06] validation step that we are doing for [27:08] the people we interact with and the [27:09] assets we're onboarding. [27:11] So it's really about like pushing what [27:14] you know and trying to one, evaluate who [27:17] you're interacting with, and two, if you [27:19] have the resource, trying to help them [27:21] to scale up their security. [27:29] Okay, thank you.