If a crypto product has never exposed its users' private keys, does that mean it is secure enough?
The answer is not that simple.
From late September through the end of October, the crypto industry was hit by an intense wave of security incidents. Major centralized exchanges and established hardware wallet manufacturers alike disclosed breaches with far-reaching consequences.
The same reassurance appeared repeatedly in the official statements that followed: “There is no evidence that private keys were compromised,” or “Users’ private keys and seed phrases were not affected.”
But attackers may not be targeting private keys at all. Exchange privileges can be abused, third-party security products can become stepping stones into management systems, hardware devices can be fitted with malicious modules, and data breaches can expose users’ real names, contact details, and home addresses. Simply saying that private keys were not leaked is no longer enough to account for the full range of risks users face.
Private-key security is only the starting point of crypto-asset security—not the finish line.
This raises a fundamental question: in today’s crypto world, how can we tell whether a product is genuinely secure and reliable?
1. How Assets and Data Can Be Compromised Without Stealing Private Keys
As discussed in the previous article, recent security incidents reveal a clear trend: attackers do not necessarily need to break cryptography or compromise private keys directly. Instead, they may look for a weak point anywhere along the broader “chain of trust.”
Employee endpoints, third-party security products, software supply chains, service containers, and transaction authorization workflows can all become points of failure. A weakness in any one of them can undermine an otherwise robust security system:
- In the Bybit incident, attackers used social engineering to compromise a Safe developer’s device, obtain access credentials, and take control of front-end resources in Safe’s wallet infrastructure.
- In the Bitget incident, attackers exploited a vulnerability in a third-party security product, obtained highly privileged internal credentials, and bypassed existing risk controls with forged withdrawal instructions.
- In the more recent Binance incident, the attack may have begun with a compromised DevOps engineer’s laptop. Binance later said that its investigation had identified additional points of compromise and that the full attack sequence was still being established.
The three incidents involved third-party wallet infrastructure, a third-party security product, and an internal DevOps environment, respectively.
This also explains why cold storage and multisig safeguards are not silver bullets. Attackers often exploit trusted components that can influence final transaction execution but have not been independently verified to a sufficient degree.
In a centralized asset-management system, attackers may not need users’ private keys. If they can control transaction instructions, approval workflows, or wallet execution systems, they may be able to trigger unauthorized asset transfers.
Beyond the fund-transfer workflow, there is another risk that ordinary users often underestimate: data breaches.
In May 2025, Coinbase disclosed an insider data-theft incident. Attackers bribed some overseas customer support personnel and used their legitimate work permissions to access internal support systems, stealing personal information, account data, and transaction-related information belonging to some users.
Coinbase said that user passwords, private keys, and funds were not directly compromised. Instead, the attackers intended to use the stolen data for targeted social engineering, impersonating official support staff to trick users into transferring their assets.
The Keystone incident further showed how a data breach could undermine device authenticity verification for hardware wallets.
On October 20, Keystone’s official website was accessed without authorization. According to Keystone’s subsequent disclosure, attackers exploited a vulnerability in a third-party website monitoring tool and obtained some customer purchase records, along with WebAuth keys used for device authenticity verification.
Even if users' seed phrases and private keys are not directly exposed, attackers could potentially use stolen verification keys to create counterfeit devices that pass official verification, leading users to unknowingly use compromised hardware.
The Ledger incident also underscores the need to protect devices from substitution, tampering, and malicious hardware implants throughout production, authentication, shipping, and sale.
These problems occur in different parts of a product, but they can converge on the same outcome: users lose control of their assets.
A private key not being directly exposed does not mean the product’s wider security system has remained intact—especially as rapid advances in AI continue to amplify these risks.
2. How Can You Tell Whether a Crypto Product Is Secure and Reliable?
These incidents bring us back to the question users actually face: what should they look for when choosing a crypto product they can trust?
Beyond brand recognition, product features, and marketing claims, imToken believes that at least three dimensions deserve priority.
1. Private-Key Control: Who Actually Controls the Assets?
This is the most basic—and most important—question when evaluating a crypto product’s security architecture.
Start by distinguishing between two fundamentally different models: custodial and non-custodial.
On a centralized exchange, users typically manage assets through an account system. After a deposit, the assets enter wallets controlled by the exchange. The platform records user balances in its internal account system and manages withdrawals, risk controls, and funds.
This model is convenient because users do not need to manage private keys themselves. But it also requires them to trust the platform’s custody arrangements, internal permissions, and operations. If the platform suffers a system failure, a compromise of internal permissions, or problems with fund management, users may be unable to access or withdraw their assets.
A non-custodial wallet works differently.
When users create a wallet, the private key is generated locally and kept under their control. The wallet provides key management, transaction construction, and signing tools, while the actual asset records remain onchain.
Structurally, the two models are fundamentally different:
- Centralized custody: User → platform account and permission system → platform-controlled wallet → onchain assets
- Non-custodial wallet: User controls the private key → wallet provides management and signing tools → user directly controls the onchain address
A non-custodial model reduces the platform’s role as an asset custodian and limits the extent to which users must depend on a centralized platform to control their assets.
This does not mean non-custodial wallets are absolutely secure.
Users must back up their seed phrases, protect their devices, and take responsibility for every transaction and signature. Assets can still be lost if a seed phrase is exposed or the user is tricked into signing a malicious transaction.
Before choosing a wallet, users should answer at least three questions: Is the private key generated on the user’s device or on a server? Are the private key and seed phrase ever uploaded to the service provider’s servers? Does the user truly control signing and have the ability to restore access to their assets using another compatible wallet?
2. Data Privacy: How Much Information Does the Product Collect About You?
The second dimension is privacy protection, which many users overlook.
Using crypto products can leave behind significant amounts of information, including email addresses, phone numbers, identity-verification data, device information, purchase records, delivery addresses, account balances, and transaction histories.
Not every product needs to collect all of this data.
Centralized exchanges generally require identity verification because of their account systems and compliance obligations. A basic non-custodial wallet, by contrast, manages keys locally and initiates onchain transactions. It does not need to know a user’s real-world identity in the same way.
The key question is whether the product follows the principle of data minimization.
For example, does it unnecessarily link a user’s real identity, contact details, device information, and onchain address?
When this data is stored centrally, attackers can more easily build detailed user profiles, increasing the risk of targeted coercion—the threat commonly described in crypto as a “$5 wrench attack.” Even the strongest cryptography cannot help if an attacker can locate the asset holder directly.
Further reading: “A Turbulent Time for Web3 Wallets: How Crypto Security’s “Sword and Shield” Are Evolving in the AI Era”
3. Security Architecture: Are Its Safeguards Continuously Tested?
The third dimension is the product’s underlying technical security capability.
For software wallets, users should focus on private-key generation, encrypted storage, transaction signing, and code security. For hardware wallets, they should also examine the secure chip, firmware integrity, device authentication, and supply-chain management.
Regardless of product type, several criteria apply broadly.
First, can the core code be independently reviewed? Open-source code does not guarantee security, but it allows external developers and security researchers to inspect critical components.
Second, has the product undergone independent security audits? Users should examine which security firms conducted the audits, which modules and versions were covered, and whether the identified issues were resolved.
An audit only assesses a defined scope at a particular point in time. It does not guarantee that the entire product is—or will remain—free of vulnerabilities.
Third, does the product maintain a bug bounty and continuous-testing program? Security is not a one-time task. As a product evolves and integrates new chains and third-party components, its attack surface changes. Public vulnerability-reporting channels, bug bounties, version reviews, and sound permission management are all meaningful indicators.
Finally, does the product provide sufficient security transparency? After an incident, timely disclosure of the impact, cause, recommended user actions, and remediation progress is an important measure of whether the product deserves long-term trust.
Ultimately, security audits, open-source code, hardware isolation, and bug bounties are not silver bullets on their own.
A genuinely reliable security system should contain risk as locally as possible, preventing a failure in one component from escalating into a complete loss of control over users’ assets.
3. How Web3 Products Can Build Security for the Long Term
For wallets, imToken’s security philosophy is to keep users in control of their assets while using product architecture and continuous security review to reduce the exposure of private keys and sensitive information.
1. Non-Custodial Architecture: Keep Private Keys in Users’ Hands
Unlike platforms that centrally custody user assets, imToken uses a non-custodial wallet architecture.
When a user creates a wallet, the private key is generated locally and kept under the user’s control. imToken does not store users' private keys or seed phrases on its servers, nor does it use a centralized account system to control onchain assets on their behalf.
The imToken software wallet generates private keys locally using cryptographically secure random-number sources provided by iOS and Android. Its core library, TokenCore, has been open source since 2018.
The imKey hardware wallet uses a true random number generator (TRNG) built into its secure chip to generate seed phrases and private keys, providing an independent hardware security foundation for key generation.
The value of this design is not simply that the user possesses a seed phrase.
More importantly, it reduces the service provider’s need to access or centrally manage user private keys. This lowers the risk that failures involving platform servers, insider permissions, or centralized custody systems will directly affect users’ control of their assets.
2. From Data Privacy to Code Audits: Security Extends Beyond Private Keys
Non-custodial architecture addresses asset control, but wallets must still manage security challenges arising from applications, servers, third-party components, and external interactions.
Regarding data privacy, imToken has stated publicly that it uses development tools such as Sentry and Firebase for error reporting and analytics but does not collect sensitive data related to wallet security. In other words, operational telemetry must remain clearly separated from security-critical information such as users’ private keys and seed phrases.
For technical review, imToken has long worked with professional security firms such as SlowMist to audit software wallets, hardware wallets, and smart contracts.
imToken also maintains a dedicated internal security team that reviews code permissions, version updates, and third-party dependencies, and conducts black-box and white-box testing before releases.
Together, these mechanisms form an ongoing security-management system covering key generation and storage, code updates, external audits, vulnerability discovery, and remediation, rather than relying on a single technology or one-time certification.
3. Security Is Not About Avoiding Every Attack—It Is About Making Users' Assets Harder to Compromise
For users who hold substantial crypto assets over the long term, relying only on an everyday internet-connected device may not be the best approach.
In addition to hardware wallets, imToken supports QR code-based transfers of transaction information between an offline wallet and a watch-only wallet.
Users can install imToken on an old phone, disconnect it from the internet, and then create a wallet while keeping the device offline. They can then connect it to a watch-only wallet on another internet-connected device through a QR code connection that supports ERC-4527.
When transferring assets, the connected device constructs an unsigned transaction and passes it to the offline device by QR code. After verifying the transaction details, the offline device signs the transaction with its local private key and passes the signing result back via QR code. The online device then broadcasts the signed transaction.
Throughout this process, the private key never leaves the offline device or becomes exposed to the internet. This adds a layer of physical isolation to day-to-day asset management and reduces the chance that a network attack can reach the private key directly.
For users seeking additional hardware-level protection, the imToken ecosystem also includes the imKey hardware wallet.
imKey uses an Infineon SLE78-series secure chip built on the Common Criteria EAL6+–certified M7893 security controller platform. It supports secure storage, secure boot, and true random-number generation.
imKey also provides a device binding code, a first-time activation process, and activation-status checks to help users determine whether the device is in the expected state.
These safeguards are especially relevant in light of the recent Ledger and Keystone incidents. Hardware security depends not only on a reliable chip but also on the device’s source, initialization process, and authenticity verification.
Device authentication is not an absolute guarantee against counterfeiting. Users should still buy through official channels and carefully verify the device’s status before first use.
Final Thoughts
The Binance, Coinbase, Ledger, and Keystone incidents show that the security boundaries of crypto products are becoming increasingly complex.
Users must protect more than seed phrases and private keys. Identity information, purchase records, device authenticity, and sensitive data exposed throughout the user journey also matter.
Before choosing a crypto product, ask yourself three questions:
- Who controls my private key, and do I truly retain control of my assets?
- What data does the product collect about me, and is any of it collected unnecessarily, retained for too long, or shared with third parties?
- How does the product demonstrate its security? Does it provide reviewable code, independent audits, and ongoing security and incident-response mechanisms?
No product can promise absolute security.
But a crypto product worthy of long-term use should reduce unnecessary trust dependencies in its architecture, limit user exposure at the data layer, and remain subject to continuous technical review and verification.
The ultimate goal of crypto security is not to convince users that a brand will never encounter problems. It is to ensure that users remain in control of their assets.