DSS 2026 — Web3 Security - Real-Time Monitoring Akand Sitra Source video: https://www.youtube.com/watch?v=_cATIJH_R5w Automatic captions — not human-reviewed. Complete spoken-audio coverage is unverified. Source language tag: en (English). Source export SHA256: 31daba992c1a7afc765f5a70d80135f530cc043d25ee277e20621172326854a1 [0:01] Hi all. My name is Akhan Citra. I'm from [0:03] Hypernative. We are a web three security [0:05] firm. [0:06] What we do is something called real-time [0:08] monitoring on 80 plus blockchains. Today [0:11] we'll talk about what exactly we do. We [0:14] try to predict risk and try to [0:16] neutralize them before they attack. [0:19] But how do we do this? [0:21] So let's look at the blockchain itself. [0:22] A blockchain is a chain of blocks. What [0:24] we do is at every block we look at all [0:28] its contents. We look at its financial [0:30] activities. We look at non-financial [0:32] activities like [0:34] what smart contracts are being deployed. [0:36] We we try to analyze the bytecode of [0:38] these smart contracts. We try to see [0:40] what functions are being called. [0:41] Sometimes you know delegate call [0:43] functions are being called. Sometimes [0:45] events are being emitted like change [0:46] admin events. We also look at different [0:48] sorts of [0:49] financial activities like [0:51] a transaction goes through. We look at [0:53] the source of funds. We look at the [0:54] destination of funds. So we monitor all [0:57] this at every block. So on Ethereum it [0:59] would be every 10 seconds, Solana every [1:02] 1 second and what not. [1:04] Then we try to detect stuff. [1:06] We run our own ML models. We have [1:09] battle-tested models which which have [1:11] been trained on past hacks and exploits. [1:13] We try to analyze whether a newly [1:16] deployed smart contract is going to do [1:18] an attack or not. We look at different [1:20] sorts of heuristics. We look at fishing [1:23] attacks. We look at scam networks. We [1:25] look at poisoned addresses and what not. [1:27] We perform a lot of simulations at every [1:29] block and these detections we also [1:32] combine with our customers' own [1:33] detections. Sometimes a vault provider [1:36] would say that I want to look at [1:40] the utilization rate of my vault. So [1:42] they set up these custom rules. We check [1:44] for these rules at every block. [1:46] So once we are detecting all these [1:48] attack vectors we are able to find out [1:51] what is malicious. [1:53] The customer are Hypernative customers [1:56] have a rule, have a playbook. They set [1:57] up a policy. When something malicious is [2:00] they want a set of things to be done. [2:02] For example, [2:04] I'm a protocol. If my contracts are [2:07] mentioned somewhere on the blockchain on [2:10] a newly deployed smart contract [2:11] maliciously, [2:13] I would say I would want to [2:14] automatically pause my contracts or, you [2:16] know, just send a notification to my [2:18] Slack channel or a Telegram channel or [2:20] there's a hack coming in. At 2:00 in the [2:22] morning, I want to get a call using our [2:24] pager duty integrations and whatnot. [2:26] Some of our asset managers say that I [2:29] have a position in a in a Morpho vault. [2:31] When, say, something malicious happens [2:34] or some financial threshold is breached, [2:36] some invariant is breached like the [2:39] utilization rate or the [2:40] collateralization ratio or there's a [2:42] depeg going on, I want to automatically [2:44] unwind my position. I want to withdraw [2:47] my funds to cold storage. I want to, you [2:49] know, do some sorts of circuit breakers. [2:51] So they set up these rules and what we [2:53] do is at every block we look at for [2:55] these maliciousness or these attack [2:57] vectors [2:58] and in the same block or in the next [3:00] block we try to execute these on chain [3:03] directly. [3:05] Many of our customers have used this [3:06] successfully. One example is [3:10] on the Flare blockchain. We have a [3:12] customer called Kinetic. [3:13] So Flare blockchain, we were monitoring [3:15] the the blockchain at every block. We [3:17] saw that a new smart contract was [3:21] was deployed maliciously and it was [3:24] targeting the Flare blockchain. So the [3:26] attacker had a very complicated logic in [3:28] mind. So he wrote the entire code on a [3:31] on a smart contract, deployed that smart [3:33] contract. Usually you need to add a bit [3:35] of gas fees to it. Then you need to [3:37] execute the transaction. It takes a few [3:39] blocks. [3:40] While the deployment happened, before [3:44] the attack happened, we detected that [3:46] this is a malicious smart contract [3:48] trying to target our kinetic [3:50] the protocol [3:52] and immediately we were able to alert [3:53] them and pause their markets. [3:55] So it it took around 4 to 5 minutes. In [3:58] this case, it took 5 6 minutes and 5 6 [4:01] minutes is like almost 200 blocks. So we [4:03] were able to trigger this transaction, [4:05] save our customer funds in real time. [4:08] And it's a $5 million save. We have [4:10] dozens of such kind of [4:12] answers as well. [4:14] So our real-time monitoring and [4:15] detection technology is used in multiple [4:18] solutions in in multiple use cases. The [4:21] our flagship product is the on-chain [4:23] monitoring and the automated on-chain [4:25] response product which is saving [4:26] customer funds. But we also have for [4:29] wallet operations. We perform [4:31] pre-transaction simulation using our [4:33] transaction guard product. We also look [4:35] at fraud detection, we look at scam [4:37] detection, we look at for regulatory [4:40] compliance, we look at [4:41] reputation of an address before you [4:43] interact with them. We look at wallet [4:44] protection for users. Not just that, [4:46] toxicity of a liquidity pool. We look at [4:49] token reputation, we look at different [4:51] sorts of attack vectors. And as on [4:53] today, we have 350 plus customers with [4:58] the leading financial institutions, the [5:00] the top the top protocols, banks, and [5:03] also top blockchains users. [5:05] We are here throughout the entire week. [5:07] You can always reach out to us if you [5:09] want to understand more about this new [5:11] technology which is actually saving [5:12] funds, the real-time monitoring and [5:16] and automated saving of your on-chain [5:18] infrastructure. Thank you.