On October 29, 2025, the on-chain security company Blockaid disclosed that the zkLink cross-chain bridge on Ethereum is currently under an ongoing attack that has yet to be resolved. The targeted components are the L1 ETH custody contract and the ERC20 cross-chain bridge related to zkLink Nova. The attacker injected forged batch roots into the Ethereum network via public relayers, successfully leading the custody contract to accept and confirm these fake data, which then executed withdrawal logic intended only for legitimate requests. As a result, approximately $786,000 worth of ETH and various ERC20 assets were transferred out of the L1 ETH custody contract and the ERC20 cross-chain bridge to an address controlled by the attacker. Both the security company and several media outlets uniformly state that "the attack is still ongoing," and no verifiable official resolution plan or compensation commitment has been announced. As a "trust hub" that carries asset and state migration in a multi-chain ecosystem, cross-chain bridges have long been seen as security hotspots. This time, the focus is no longer on the accidental leakage of a key but rather on the trust cracks within the public relayer and batch root verification processes. This article will delve into this crack and raise necessary inquiries about the structural risks of cross-chain bridge verification and relaying mechanisms.
The attack is still ongoing: $786,000 has been transferred out
Beyond the technical details, this incident is initially quantified into a number that gives an intuitive chill. On October 29, 2025, when Blockaid issued a warning, it pointed out that the attack was no longer in the tentative phase but had indeed resulted in multiple illegal withdrawals. According to AiCoin data, by the time of the disclosure, approximately $786,000 in assets had been transferred out from the zkLink Nova-related L1 ETH custody contract and ERC20 cross-chain bridge through both channels, directly flowing into the attacker-controlled wallet address. The affected parties include not only users relying on the "safest" Ethereum L1 custody layer but also the ERC20 bridging contracts that handle daily cross-chain liquidity. Both trust links that should have supported each other were torn apart by the same type of forged batch root all at once.
What continues to exacerbate the tension is not just this figure itself, but the "ongoing" state behind it. Blockaid explicitly described the incident as an ongoing attack in its announcement, and numerous reports cited this description rather than declaring a clear endpoint. In the absence of verifiable official resolution plans, contract suspension announcements, or compensation commitments, users find it difficult to determine whether the losses have been capped, whether new forged batch roots will still be accepted by the bridge, and where the security boundaries of custody and cross-chain channels truly lie. Current public information mainly comes from Blockaid's on-chain monitoring and media reports. For those holding tokens using the zkLink cross-chain bridge, such a singular information landscape and ongoing attacks are themselves a collective stress test of the trust structure within the multi-chain ecosystem.
Public relayers breached: How forged batch roots penetrated Ethereum
In day-to-day operations, cross-chain bridges rely on public relayers and batch roots to complete a "multi-chain bookkeeping." Public relayers are responsible for packaging cross-chain requests and balance changes accumulated from various chains and reporting them, while the L1 custody contract on Ethereum verifies whether these aggregated states are trustworthy through batch roots. Batch roots serve as a compressed summary of a cross-chain bookkeeping cycle, designed to allow any party to submit relay data, but only if the contract strictly verifies it will it be considered a legitimate state and trigger asset cross-chain transfers or withdrawals.
This attack path cunningly exploits this public mechanism in reverse. The attacker did not attempt to steal a key but rather injected forged batch roots into the Ethereum network via public relayers, misleading the L1 ETH custody contract and the ERC20 cross-chain bridge contracts into treating them as real cross-chain states. The more critical breakthrough point was that the custody contract performed "final confirmation" on the batch root, failing to distinguish whether it was legitimate or forged during the verification process, thereby executing withdrawal logic that should have only targeted normal cross-chain requests. The result was approximately $786,000 worth of ETH and ERC20 assets being transferred out from L1 custody and cross-chain bridge contracts into the attacker's wallet address, which also directly corroborated the industry's focus on this event's technical emphasis being attributed to the verification processes of public relayers and batch roots rather than a single key leak.
Trust assumptions of cross-chain bridges are broken: The verification layer becomes a new minefield
In a multi-chain world, cross-chain bridges have always been seen as the "trust hub" for assets and states: users lock up ETH or ERC20 assets on one chain, while obtaining corresponding bookkeeping rights on another chain. The core of the entire process lies not in the front-end interface, but in the invisible link between the relayers and L1 verification contracts. Most cross-chain bridges rely on relayers to report the aggregated state of cross-chain operations, which is then verified and "finally confirmed" by the L1 contract; once batch roots are deemed valid, the contract will execute the release and transfer of real assets at the custody end. This carries the foundational trust assumption of the multi-chain system: that the on-chain verification layer can reliably distinguish between legitimate and illegal state updates.
To increase decentralization, many projects opt for a public relayer design, allowing more participants to submit batch roots to reduce concerns about a single entity becoming a "black box relayer." However, the tension between open participation and security boundaries has been laid bare in the zkLink Nova event: the attacker did not steal an admin key, but directly exploited public relayers on Ethereum to inject forged batch roots, which were accepted and confirmed by the L1 ETH custody contract and ERC20 cross-chain bridge contracts, ultimately triggering illegal withdrawals of approximately $786,000 in assets. Existing materials clearly lock the technical focus on this verification link rather than a single private key leak, indicating that the risk narrative of cross-chain bridges is shifting from "signature layer breach" to "verification layer breach." For all projects adopting similar relayer and batch root architectures, this attack not only reaffirms cross-chain bridges as a security hotspot but also exposes the verification layer itself as a new minefield, forcing the industry to reevaluate how much security redundancy can still be retained between open relaying and trusted verification.
From zkLink to the entire industry: A long-standing issue of frequent bridge attacks
If we place the zkLink incident back into a longer timeline, cross-chain bridges have long been a section where security companies and the tech community have repeatedly sounded the alarm. Multiple significant on-chain security incidents have erupted in cross-chain scenarios, often not concerning the final asset contracts but rather revealing design gaps along the collaborative links between relayers, validation nodes, and custody contracts. Security research and industry reports thus commonly classify cross-chain bridges as high-risk infrastructures within the multi-chain ecosystem: they serve as the "trust hub" for cross-chain assets and states, while also stacking complex logic at the verification and relaying layers, making them the preferred entry point for attackers to repeatedly probe.
In such a lineage, the zkLink attack is not an isolated exception but continues the old problem of "weakness in the verification layer and relaying layer": the combination of public relayers and batch root verification was breached, ultimately causing the L1 custody contract to mistakenly trust a forged state and execute withdrawal logic that should have only targeted legitimate requests. In the long run, such incidents directly erode users' trust in cross-chain bridges, making "whether cross-chain is safe" a re-emerging constraint on the pace of multi-chain expansion, and forcing teams to allocate more resources to verification models, relaying operations, and contract auditing, while making more cautious trade-offs between cost and security.
What signals should users and project parties focus on next?
At the juncture around Blockaid's disclosure of "the attack is still ongoing," ordinary users should not guess technical details but should focus on several types of public signals. Firstly, the information channel: if zkLink subsequently issues announcements on its official website, social media, or documents, it's crucial to note whether it specifically names the affected zkLink Nova L1 ETH custody contract and ERC20 cross-chain bridge contract, whether it provides clear risk levels and operational advice such as "suspend use / migration paths." Secondly, on-chain status: users can observe via browsers whether the relevant custody contracts exhibit actions such as owner calls, permission changes, contract logic upgrades, and whether the attacker's address, which previously received approximately $786,000 in assets, continues to obtain illegal withdrawals from these contracts; these are all intuitive signals to determine if the attack has truly been halted. The actual risk to users largely depends on whether they continue to cross-chain through the affected bridge and whether they have received sufficiently clear alerts and operational guidance from the public information.
From the perspective of project parties, the common path following cross-chain bridge attacks within the industry has become relatively clear: urgently suspend the relevant bridge contracts, collaborate with security teams including Blockaid to ascertain the nature of the event and trace it on-chain, attempt upgrades to verification logic and relaying architecture, and discuss fund disposition and compensation plans when feasible. Current public information has yet to reveal a verifiable official resolution plan from zkLink, thus, the most noteworthy variables to observe next will be whether zkLink presents a technical review over an extended timeframe, explaining the specific breach points in public relayers and batch root verification; whether new verification and relaying designs emerge on-chain, responding to inquiries with real contract and parameter changes; and whether other cross-chain bridge teams proactively adjust their verification and relaying models, using this incident as a common starting point to reassess security boundaries. Only with gradually emerging clear actions and structural changes on these dimensions can multi-chain users redefine the trust boundaries of zkLink and the broader cross-chain bridge ecosystem.
Join our community, let’s discuss and grow stronger together!
AiCoin exclusive Hyperliquid benefits: https://app.hyperliquid.xyz/join/AICOIN88
AiCoin exclusive Aster benefits: https://www.asterdex.com/zh-CN/referral/9C50e2
On-chain Telegram community: https://t.me/AiCoinWhaleData
On-chain community: https://www.aicoin.com/link/chat?cid=N6OVMor5g
AiCoin on-chain Twitter: https://x.com/aicoinwhaledata
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
