Dialogue with Shenyu: My personal experience of having 12,000 ETH stolen and the role of AI in Web3

This discussion started with the $1.5 billion theft from Bybit, and focused on the security vulnerabilities of multi-signature wallets (such as Safe) and their solutions. Shenyu pointed out that the front-end, hardware, browser and other infrastructure that multi-signature wallets rely on have weak links, especially the front-end tampering and blind signing problems, which lead to inconsistencies between transaction intentions and actual operations, and are easily exploited by hackers. To this end, he proposed temporary solutions such as domain name whitelists and transaction parsing plug-ins, and advocated an end-to-end closed-loop risk control system, combining AI with third-party verification to improve security. In addition, Shenyu shared for the first time his personal experience of being phished for 12,000 ETH last year, emphasizing the risks of blind signing of hardware wallets, calling on the industry to adopt a layered and decentralized, zero-trust architecture, and strengthen the construction of a security culture. He also mentioned that in the face of state-level hacker attacks, the industry needs to respond to challenges through technology iteration and security awareness enhancement. Finally, Shenyu looked forward to the prospects of the combination of AI and Crypto, and believed that AI Agents may play an important role in blockchain networks in the future and promote industry innovation.

Reflecting on the vulnerabilities of multi-signature-dependent infrastructure, we proposed temporary solutions such as domain name whitelists and transaction parsing plug-ins, and advocated end-to-end closed-loop risk control

Colin: Please share your views on the Bybit incident, and what points you think are worthy of attention in this incident.

Shenyu: This is actually a very typical situation. Because the current industry solution for multi-signatures needs to rely on a lot of infrastructure and intermediate services, which are developed by multiple companies. In the traditional sense, this is more decentralized and decentralized, and it is not easy to have problems, but the problem now is that there are some bugs in the interaction between the applications developed by these multiple companies and the hardware technologies, and there are some possible weak points. As a result, a series of large-scale security incidents have occurred in the past period of time, facing the national-level hacker force. The essence of this problem is that I have been reflecting on this problem since I was attacked in September last year. We also contacted various hardware manufacturers around November. We realized that blind signing is a very serious problem, and then the connection from the front-end to the desktop plug-in to the hardware is also easy to be tampered with. We realized that this is the weakest point. At that time, we also contacted OneKey, Ledger and other companies to discuss some solutions. In this process, some problems were found, and each company had some solutions, but it was difficult to finally implement this solution and resist the attack. As someone mentioned earlier, it may take more than half a year to iterate, and Ledger actually has a set of systematic solutions. Because of the contract, they need to be updated passively, and this cycle is also very long. How can this full link better plug the loopholes or cross-produced bugs?

The market is actually lacking an end-to-end solution. Now, it is combined by various companies, but there will be accidents in the process of combination, which will give hackers an opportunity to take advantage. In this process, I actually said that we first made some internal small tools and some demos. First, we made a whitelist of domain name access to ensure that the website opened by the web page is correct and not tampered with, which can prevent some typical phishing attacks, especially things like wrong URLs and random web page jumps. Then, secondly, we made a transaction parsing plug-in that can be run on the mobile terminal. Even some hardware wallets and plug-ins or Safe communicate through QR code transmission. We will verify whether the QR code transmission has been tampered with, and then verify the parsed content on the hardware wallet. We made some small plug-ins, but I feel that they are too scattered when used. The entire end-to-end process is not completely connected, and there are too many steps to use. So after this incident, we continued to reflect.

The core point is that our industry is now very large, with trillions of dollars, which is bound to attract attacks from high-level hacker teams. In this process, as our team members also mentioned, you actually have to dig very deep both horizontally and vertically. However, because the industry is indeed developing very fast and the iteration is very fast, people often ignore this series of potential risks when doing business. So in this process, our current idea or what we are doing is that we hope that because we have been managing various private keys (private keys on hardware and software chains), we have accumulated a series of risk management capabilities and some risk control engines in this process. So we hope that in a typical scenario like Safe, we can be a custodian, we take a private key, and under the condition of taking a private key, we have a completely independent software and hardware environment, plus our series of risk control engines to analyze, and then at the same time introduce our customized series of audit solutions in this process, plus automated AI analysis, plus manual audit, and then add some black and white lists to it, and even some advanced contract parameter control things.

This is actually something we have been using in the process of DeFi, but we have not completely linked it up for productization. Through this form of decentralization, some private keys are not obtained by a team, but by some external third parties independently, so that this matter can be closed end-to-end and controllable. This is our current idea. We do this in the process of operating DeFi on the chain, because EOA is indeed very easy to be phished, and migrating to multi-signature faces problems similar to Bybit. We have a very long link and various risks. Our current thinking and solution is to introduce an independent third party, and then the independent third party introduces its completely independent technology stack and integrated hardware and software solutions, including risk control engines, and even adds some AI capabilities to complete a closed loop, from the initiation and analysis of transactions, risk control review, to the signing process, and then try to avoid the very patient and long-term penetration attacks of national hackers.

Shenyu’s experience of being phished emphasizes the risk of blind signing with hardware wallets, and the need to combine AI and third-party verification

Colin: You mentioned the issue of EOA phishing just now. We also know that some of your assets may have been accidentally phished last year. Can you recall what happened at that time and whether the money was finally transferred by the so-called hackers from North Korea?

Shenyu: My background at that time was that a certain project was airdropping. My physical condition was not so ideal at that time, and I was a little distracted, so I clicked on the wrong link. There was a link with a problem. But the problem is that the funds have reached the hardware wallet side. In fact, we have a three-party mechanism for risk control of domain names and DNS resolution. As a result, that solution was bypassed, and it happened that our risk control did not hit. After bypassing, I was a little distracted, so I didn’t check it so carefully, and then it reached the hardware wallet side, because the hardware wallet side is blind signed. Then after I pressed that, I felt that there seemed to be something wrong. I immediately went to check and found that it was wrong. The rest is what everyone knows. So after this happened, we went to solve the problem of blind signing of hardware wallets. In this process, during the National Day last year, we held meetings with OneKey and other parties, and found that the problem was not so easy to solve. Because EOA is easy to be phished, especially easy to be targeted and "attacked". Then we turned to using Safe for multi-signature. During the multi-signature process, I found that this problem was somehow more serious, because basically every transaction was blindly signed, and then we had to make a lot of small tools to try to solve these problems. In the end, there is still a need for a closed-loop solution, and our hardware wallet needs to achieve the goal of hardware and software integration, because the hardware UI is indeed the last point to check, and some independent third parties need to be introduced. In this process, it can prevent interception, alarm and processing when the human state is not right, so this is why we started iterating in this area and trying to productize it.

The reason why the “copycat season” did not appear: lack of driving force and expectation that the national reserve decision will promote market development

Colin: In addition, last year you first raised the question of "altcoins" and everyone argued fiercely. Some people, including many well-known people, criticized that there must be "altcoins", and some people also accepted it. Then in December last year, the "altcoin" market was indeed very short-lived. At that time, you may have felt that the "altcoin" market was coming and said that the "altcoin" market had already begun. As a result, not long after, it seemed that "altcoins" were almost non-existent in this cycle, as you said at the beginning. Of course, we are not making predictions. As for short-term predictions, I don't know who said it before, only God knows. But now do you have some new thoughts and feel that it is almost impossible for "altcoins" to appear in this cycle, because it may be mainly in the market cycle of Bitcoin. Also, don't you think that the so-called bull market has ended, or that it is about to turn into a bear market?

Shenyu: My current feeling is that in the past two or three years, apart from some small emotional hot spots, the entire industry still lacks a very clear application and real demand scenario like in 2020 and 2021. I think the essential problem is here. Because there is no endogenous driving force, there will be no new truly valuable application assets settled. On the other hand, in this cycle, a large number of players actually stayed in the traditional US stock market. They traded through ETFs and platforms such as Robinhood. They did not really own cryptocurrency assets, so a lot of money did not stay in the market, and there was no so-called spillover effect that everyone wanted to look forward to, starting from Bitcoin and Ethereum to other currencies. With these two factors combined, somehow, there may be a "coltcoin" market that is also very short-term emotionally driven, just a few weeks, and there is no general outbreak. I am almost still maintaining my previous judgment.

Then my view this year, or my expectation for the market, is that I think the market may develop better in the second half of this year, maybe from June to October, and then after the US national reserve issues become clearer and passed, the entire industry or market may see more new funds flowing in. But now, at this stage or in the short and medium term, we may not be able to solve the problem in the application level in the short and medium term. There doesn't seem to be such a crazy inflow of funds at the on-site and off-site levels. So I am more looking forward to the second half of the year.

I will not judge this matter today. It may ultimately depend on whether the national reserve-related issues in the United States can have some results this year. If there are no results, the market may be over. At this moment, we think the probability of passing is still relatively high, and it is hard to say, so more expectations are placed on the second half of the year. (This Space release time is February 25, and Trump has signed an executive order on Bitcoin national reserves in March)

Summary of historical theft cases: Dealing with state-level hackers requires layered decentralization, zero-trust architecture, and security culture penetration

Colin: Actually, Shenyu, you are also a veteran in the cryptocurrency circle. I joined in 2017. There have been so many thefts in the cryptocurrency circle, all kinds of thrilling. Of course, the amount of money stolen from Bybit this time set a historical record, but it is profitable enough, so it can be fully compensated. In your historical memory, including the early days, what are the most profound theft experiences you remember? Which ones are the most worth sharing?

Shenyu: I think the attack and defense have been upgrading. The earliest attack methods were very primitive. We practitioners must realize that we are facing a national force. They are not ordinary hackers. They are organized and have been intensively trained since they were in their teens. They use various methods and methods similar to attacking core facilities to infiltrate our companies. They will also challenge human nature. Everyone must clearly realize that we are facing such an opponent. In this process, people will relax and there will be some problems with human nature. In the end, we must adopt means and methods that can withstand it.

In the history of the Internet, Cobo may be the first company with a Chinese background to adopt the zero-trust model, a security methodology. The reason why we adopt this methodology is that only this methodology has been verified to withstand the infiltration of state-level forces. So we started to promote zero-trust transformation internally as early as 2018 and 2019, and we had to install various things on all our internal services and all employees' computers and mobile phones. So after we realized this, we must adopt this solution and put our various systems in a state of minimum trust.

At the same time, the most core asset for us is the private key, and we must introduce the thinking of layering, decentralization and dispersion. What does layering mean? We must split the wallet very loosely. I have personally shared my four-wallet theory before, which is at the personal level. But at the institutional level, there must be at least a three-layer wallet architecture of hot, warm and cold, and each layer of the architecture may have its own characteristics, and there may be black and white lists and a series of processes, including time delays. Many times, we often compromise and sacrifice security for efficiency, especially in this industry that is developing very fast, but often the mandatory time delay will reduce your security risk exposure very low, especially at the level of cold wallets and warm wallets. We must layer, and then set different security risks for each layer, and establish a set of audit systems and processes to avoid these systemic risks. Cold wallets are best at the physical level, which can guarantee absolute security.

Then the second is to decentralize. Because the industry has developed for so many years, there are more and more participants. At first, we may not have a solution, and we can only trust our team to do something. Now, internal employees, especially in the case of popular remote work, have also been infiltrated by North Korea, and even high-authority employees have been obtained within the organization, so we cannot completely trust this set of internal processes. Therefore, decentralization is necessary at this level. We must introduce an external independent third party to control some private keys and verify them, which is also very critical. Now there are many custodians, security companies, and insurance companies. They can hold private keys at different levels, some of which are hot wallet private keys, some are warm wallet private keys, and some are cold wallet private keys. As an external independent third party, they have their own security solutions, and then do some risk management and control. Through a series of measures, the cost and threshold of attack will increase exponentially.

Then, in the end, it is actually about decentralization. We may have done a good job in distributing our software and hardware around the world, because everyone is basically in a decentralized state. From this dimension, first of all, we must use a minimum trust system and the zero-trust risk design concept to design our entire internal system and architecture, and then use the layered, decentralized, and decentralized management of core assets, and then equip a series of software and hardware security modules, as well as strict internal access control processes, and then establish a set of closed-loop management of the secure life cycle. In addition, we may have some emergency responses and solutions during and after the event, so that we can survive for a longer period of time in high-risk and uncertain attack events.

Compliance exchanges invest more in security, while offshore exchanges are under great growth pressure and lack security measures

Colin: I personally want to know about Coinbase or other compliant exchanges. To be honest, it seems that there are fewer thefts, unlike offshore exchanges, which are almost all stolen, and even many exchanges have been stolen many times. What is the reason? In fact, I don’t quite understand. In theory, can offshore exchanges also use the architecture they can use? Because offshore exchanges have larger amounts of funds or different operations. I don’t know which security guest can answer this question for me. In addition, Shenyu, do you think that facing this kind of attack by North Korean hackers will make it very difficult to start a business in this industry, causing the security costs of ordinary entrepreneurs, or the investment they need to make, to become very high, which will greatly hinder the development of this industry? It seems that everyone has some doubts about whether the industry can withstand the attacks of North Korean hackers.

Shenyu: Let me add that my intuition is that compliant exchanges may pay more attention to security than efficiency. For example, they are stricter in some aspects, so they also invest a lot in security.

In theory, offshore exchanges have a lot of funds and can invest more in security. Whether it is the early Binance or others, including this case, it feels that thefts are still quite frequent.

It may be because the offshore exchanges are under too much pressure to grow, and they need to iterate at a high frequency, and there are many complaints from users. However, in compliant exchanges, ordinary users do not have such high expectations, and withdrawals are not required to be so fast. Even large withdrawals are sometimes T+1, T+2, and I have even seen T+7, and users can accept it because the customer base is different. The customer base of compliant exchanges is mainly institutions. These two exchanges may have a long history. I don’t know if they have experienced thefts. They should have a lot of experience in dealing with security issues internally. Security issues can basically be regarded as the "coming of age" of the company.

I think as long as there is enough profit margin in this industry, investing in some SaaS products will definitely be able to try to meet the pain points and needs of this industry, but now people are not so aware of security or willing to pay. We have also seen many very good security products, but everyone earns money through hard work, and even the costs are difficult to balance, so they need to rely on other aspects to subsidize.

This is actually a problem, but I feel that with the upgrade of attack and defense, people are gradually realizing that security is very important, and the investment in security will increase. This also provides a certain amount of development space and funds for companies like security SaaS. From the perspective of security and architecture, there are effective and verifiable solutions, but in the case of Safe, there are about four to five participants upstream and downstream. To complete a transaction, the coordination between each participant is very slow, and the hardware iteration is particularly slow, so in the end, there is a time window for hackers.

Once the problems in the blockchain industry are fully exposed and discussed, they should be able to be solved through one or two cycles of iteration. In addition, Web2 also faces similar problems, although the resources invested in security issues in Web2 are not as large as those in the blockchain industry. We have seen that Passkey, which is used to protect password security, has been promoted for many years and has finally begun to be used on a large scale in the past one or two years, especially in some sensitive financial fields. Therefore, the underlying technologies can be reused and developed, including the Apple devices we use, which are also rapidly iterating on the security level. In the end, there is a solution, which may just require a certain amount of time and financial investment. In this process, some developers who are not very risk-aware and act more aggressively may bear some costs, but the problem can be solved.

Security advice for entrepreneurs: practice zero trust model, multi-audit cross-validation, and regular security drills

Colin: For entrepreneurs, recently a startup project was stolen 50 million US dollars, although many people in the community are also supporting them. As a person with many years of experience, you have experienced so much. What suggestions can you give to entrepreneurs to better improve their security awareness?

Shenyu: I think it is very important to practice the zero-trust model in the process of entrepreneurship. Only in the current environment, only this set of methodologies and concepts can protect everyone. At the same time, we cannot rely on some single point failures and unilateral contract audits. The basic requirement is that at least two or three companies must conduct audits, and some cross-validation is required, so that some problems can be exposed in the process. And don't raise funds too quickly. In the early stage, you can slowly raise funds through some internal testing and public testing, and isolate the funds, so that you can manage risks relatively well.

In fact, there are many inexpensive security solutions in the industry, such as security monitoring and risk control systems, which should be utilized by everyone, as they can greatly increase the probability of survival.

In the process of starting a business, on the one hand, the business model must consider the user side. Especially for entrepreneurs without a technical background, they must spend a lot of attention on security and internal zero-trust architecture, at least 20 to 30 percent of their attention. If the company does not emphasize these internal security culture and systems, and does not regularly conduct internal security phishing tests and attack and defense drills, everyone will slack off from the employee level and human level. You must know that hackers may be watching you at any time. Therefore, resources and attention must still be placed on security.

Colin: Yes, I think that in the process of this industry growing, almost no company, whether it is the boss or the company, will not encounter theft. Fortunately, as long as this time is not defeated, both individuals and the entire industry will make some progress.

Why not participate in this round of Memecoin: Feeling unwell, focusing on AI

Web3 Dumplings: Shenyu, you have been in the crypto industry for a long time and are also a well-known figure in this industry. At the earliest, Cobo Wallet may lead everyone to mine together, and I participated in all of these. Under this year's market conditions, I found that Shenyu rarely spoke about industry-related matters. Including the link sent today, the PVP mentioned was only briefly mentioned. What I am more concerned about is that for the industry, it seems that after the last wave of market conditions, the industry has reached a bottleneck. I would like to ask Shenyu, where is the next outlet?

Colin: Yes, Shenyu, you have always liked to try out all kinds of new things, but it seems that this round of Meme is so popular, but I haven’t seen you participate in any special way. What’s the reason?

Shenyu: The main reason I don’t play Meme is that my body can’t handle it, and I can’t compare with the post-2000s generation. Another factor is that I have been focusing on AI recently. AI seems to have some disruptive combination with cryptocurrency in the future, which may bring new growth. We proposed in an offline event more than a year ago that the ultimate users of cryptocurrency may be AI Agents or AI robots, not humans, so I spent more energy on learning how to use AI. I really can’t play PVP anymore. I usually participate just to give money to everyone, and there is not much positive feedback, so my main focus is on AI.

Looking forward to the prospect of combining AI and Crypto, it is believed that AI Agent may play an important role in the blockchain network in the future

Colin: What do you think of AI now? Because there was a wave of so-called AI Agents before, and they fell sharply. Many of my friends went bankrupt after speculating on AI. Do you think the previous AI agents have produced anything attractive? What do you think are the better application directions for AI and cryptocurrency in the future?

Shenyu: I think the bottom line is that AI itself has not yet reached that stage, and it is still in the concept verification stage. Then we can see that AI is developing very fast, and the underlying computing model is constantly iterating. We expect AI to reach a state in the future, that is, to have general AGI (artificial intelligence).

There are two points about cryptocurrency. On the one hand, it is very friendly to completely digital things because the data is open and transparent. So on the one hand, AI may change the way of interaction. For these complex smart contracts and anti-human security operations, we should be able to rely on some reliable and intelligent AI agents to help us make some auxiliary decisions. I like to joke that I am now operating decentralized finance and I have to arrange several AI agents to help me watch remotely. In the future, it may really be like this, with artificial intelligence engineers and maybe one or two AI agents watching the screen. Looking further ahead, AI may really be connected to the blockchain network.

We imagine a scenario where AI agents need to interact with each other, including the exchange of value and data, and may also involve things at the contract level, or even a loosely organized company or DAO. They may use a smart contract platform to exchange some value. I think that in the next three or five years, something similar to a Web social network or a value social network may be formed on the blockchain, and this will ultimately be used by this bunch of AI agents. Once this thing appears, the value and influence it brings may be very large. We have always talked about the network effect of Web3. The asset scale or value brought by this thing may be much larger than the traditional Internet companies we see now, so this may be something higher than the current trillion-dollar scale. I have been thinking about what this is and what we can do in it. I am more optimistic. Although the industry has not emerged with good application directions in the past two or three years, I think it should appear in the future, and there are still some obstacles that we may be able to solve, so I am looking forward to that day.

Safe Wallet Blind Signature Solution: Enterprise Signature Tool + AI Risk Control + Blacklist and Whitelist Management

TheCheerSong: I am a trader doing on-chain automated trading. After this incident, we are also upgrading our security measures while our business cannot be stopped. I feel that the most troublesome part is the blind signature part of the Safe wallet. What we can do now is that we have some open source permission control modules, and then we apply them to the Safe wallet. In this usage process, most of the transaction requests are automated, so the Safe wallet is basically only used for manual transfer operations of some tokens. I would like to ask all the teachers, for relatively simple requests, are there any tools to verify our signature content?

Shenyu: Actually, someone mentioned earlier that we will release this set of internal tools as a product next week, and release this set of Safe enterprise signing tools. The essence is to obtain a Safe private key, and then add some blacklists and whitelists through the machine, including some risk control templates. You can customize some common risk control measures, such as limiting the amount of tokens, transaction speed, setting blacklists and whitelists, and adding some capabilities of AI Agent, as well as considering some large losses. We can sort out this set of processes and manage risks at the same time.

This solution is combined with Cobo's previous access control list-based contracts and parameter-level controls on the Argus chain. I feel that at present, only by doing this can I feel at ease when my large funds are traded on the chain. This is the security practice we are using.

Security experts’ opinions on wallet security issues in the Bybit theft incident

During this Space, BlockSec CEO Professor Zhou Yajin, OneKey Chief Growth Officer Nig and Cobo Chief Security Officer Moon also expressed their views.

Regarding the exchange security issues in the Bybit incident, Professor Zhou Yajin said that the incident was caused by the inconsistency between the operation and the actual transaction when using the Safe contract wallet to manage funds, which led to the malicious upgrade of the wallet and the theft of funds. The reason has not been made public yet. Many project parties believe that it is safe to use Safe multi-signature to manage wallets, but in fact they ignore that security is a system construction covering operations, non-technical and technical levels. There are loopholes in private key management and transaction interpretation, such as insufficient custody, signature and transaction analysis and interpretation, and poor handling of security links brought by user friendliness when ensuring the physical security of private keys. The trust chain of large-scale multi-signature transfers is long, but there is a lack of third-party double verification of transaction interpretation and operation interface information. Therefore, the use of contract wallets to manage large amounts of funds requires the introduction of third-party verification and authentication, and a whitelist or decentralization should be set up to implement flexible policy control on contract wallets.

Regarding the frequent tampering of the front-end of the Safe wallet and the security of the mainstream multi-signature solution Safe, Moon believes that the Safe solution and the contract itself are relatively safe, but the trust chain is long when applied, and unexpected situations are prone to occur. The Bybit attack this time is most likely an outsourcing problem rather than a contract problem, which highlights the importance of improving security awareness on a daily basis. To use the Safe wallet safely, it is necessary to consider both vertically and horizontally. Each link in the vertical direction must have a controllable technical solution, such as independent equipment; horizontal multi-signatures require independent verification by the signer, and a long trust chain is prone to cross-linking, which hackers can use to bypass verification. Therefore, horizontal expansion not only requires increasing the number of signers, but also ensuring that the signer's solution and environment are independent and verified separately to give full play to the role of Safe. In addition, in addition to using contract wallets, exchanges and high-net-worth individuals should also establish strict mechanisms such as manual reconciliation and monitoring anomalies, use automated program audits, and improve security awareness to avoid Web2 attacks. In addition, Cobo will also launch a management solution with MPC plus Safe, using Safe's multi-signature capabilities to allow different signers to have independent and complete end-to-end signature links.

Regarding the hardware wallet-related issues in the Bybit incident, Nig said that Bybit used NPC to control the EOA wallet (because its signature is easy to parse), while the Safe smart contract wallet signature is complex and difficult to parse, and the security team may not have noticed it in time. Existing hardware wallets (such as Ledger) have limited performance and are difficult to deal with complex smart contract parsing and blind signing. If measures had been taken earlier, this attack might have been avoided. The Shenyu team and OneKey have developed relevant parsing tools. Ledger's Clear Signing is progressing slowly, and signature data is easily contaminated when transmitted from networked devices. It is not enough to rely solely on hardware wallet parsing to ensure consistency of intention. Bybit lacked early warning this time. After the first signer was hacked, others signed blindly due to equipment problems. It is recommended that high-net-worth individuals and institutions isolate networked devices related to fund transactions from office equipment to reduce the risk of intrusion. Previously, Radium was hacked because the environment was not isolated.

Regarding Safe security, third-party monitoring functions, and the application of AI in blockchain security, Professor Zhou Yajin believes that Safe contracts have been highly secure in the past and have been reviewed by multiple parties, but the long use process brings risks. The Falcon Safe security system developed by his team can examine user transactions from a third-party perspective, analyze transaction content, and remind key transaction information such as transfers and contract interactions, thereby lowering the user's cognitive threshold and avoiding security issues. In terms of AI applications, on the one hand, AI reduces the cost of doing evil and facilitates the batch generation of phishing tools; on the other hand, the industry is exploring the combination of AI and auditing and automated auditing codes. Although it is far from the ideal state, AI can lower the threshold for cryptocurrency users to use products and help solve complex user operation problems.

Regarding hardware wallet-related issues, Nig responded that the company would not issue a war report to promote a huge increase in sales like some of its peers after the exchange was stolen. Although the sales increase is good, it also reflects that many people have just paid attention to the security of private keys. Hardware wallets are responsible for this security incident. The performance of Ledger and Safe did not meet expectations. Safe has stopped front-end and native support. The previous generation of hardware wallets was limited in parsing capabilities due to security considerations. The new generation of products such as Ultra and Pro will strengthen local complex contract parsing and support the parsing of core transaction elements. Classic will also select key parts for display; APP will realize mainstream EVM transaction parsing, and the hardware end will be later due to security testing. For Safe, methods to resist related attacks will be demonstrated in the near future, and user safety education will be carried out. In the future, despite the diversity of technologies, such as OKX's lack of enthusiasm for hardware wallet access and institutions promoting NPC wallets, there is a risk of mnemonic exposure in the signing process in an online environment, so hardware wallets always revolve around the core of physical isolation. Even if there are changes such as mnemonic standard upgrades, the core security defense will not change.