Inside the KelpDAO Bridge Exploit: How a DVN Configuration Change Led to a $292M Loss

On April 18, 2025, the KelpDAO bridge suffered a devastating exploit resulting in approximately $292 million in losses. In response, LayerZero Labs published a comprehensive forensic report on Sunday, produced in collaboration with cybersecurity firms Mandiant, CrowdStrike, and zeroShadow. The report reveals a critical, previously undisclosed configuration flaw: the bridge’s Decentralized Verifier Network (DVN) was downgraded from a secure 2-of-2 multi‑signature setup to a vulnerable 1-of-1 single‑signature arrangement shortly before the attack. This Q&A explores the incident, the technical misstep, and the broader lessons for cross‑chain security.

What exactly happened during the KelpDAO bridge exploit?

On April 18, an attacker exploited a vulnerability in the KelpDAO cross‑chain bridge, siphoning approximately $292 million in digital assets. The bridge is designed to facilitate token transfers between different blockchains, relying on a Decentralized Verifier Network (DVN) to validate transactions. According to the forensic report by LayerZero Labs, the exploit was made possible because KelpDAO had inadvertently downgraded its DVN from a 2‑of‑2 multi‑signature configuration to a 1‑of‑1 setup. This meant that only a single verifier was needed to approve a transfer, drastically reducing security. The attacker was able to compromise that single verifier and approve fraudulent transactions, draining the bridge’s funds. The attack happened swiftly, and the stolen assets were then moved to various exchanges and mixing services.

Inside the KelpDAO Bridge Exploit: How a DVN Configuration Change Led to a $292M Loss — Source: thedefiant.io

Who conducted the forensic investigation, and what did they discover?

The investigation was a collaborative effort led by LayerZero Labs, the team behind the bridge protocol, in partnership with three top cybersecurity firms: Mandiant, CrowdStrike, and zeroShadow. Their combined expertise in incident response, threat intelligence, and blockchain forensics produced a detailed timeline of the exploit. The key finding was the unauthorized change in the DVN configuration. The report notes that the downgrade from a 2‑of‑2 threshold (requiring two verifiers to approve) to a 1‑of‑1 threshold (requiring only one) occurred shortly before the attack. This alteration eliminated the redundancy that would have prevented a single point of failure. Furthermore, the investigation traced the stolen funds and identified the attacker’s wallet addresses, though the perpetrator remains unidentified.

What is a Decentralized Verifier Network (DVN), and why was the configuration change so critical?

A Decentralized Verifier Network (DVN) is a security mechanism used by cross‑chain bridges to validate transactions between blockchains. In layman’s terms, it acts like a panel of independent inspectors who must agree before a transfer is approved. The DVN’s threshold settings determine how many inspectors must sign off. A 2‑of‑2 configuration requires two separate verifiers to confirm a transaction, providing a robust safeguard against compromise. However, a 1‑of‑1 configuration trusts just a single verifier. If that single verifier is compromised or malicious, the entire bridge becomes defenseless. In the KelpDAO case, the downgrade to 1‑of‑1 effectively turned the bridge into a single‑point‑of‑failure system. The attacker likely exploited this by gaining control of that one verifier, allowing them to approve fraudulent cross‑chain messages and drain the bridge.

How did the attacker exploit the 1‑of‑1 configuration?

Once the DVN was downgraded to 1‑of‑1, the attacker needed only to compromise or impersonate that single verifier. According to the report, the verifier was likely an off‑chain oracle or a node that held a private key. The exact method of compromise is still under investigation, but possibilities include a phishing attack, social engineering, or exploitation of a software vulnerability in the verifier’s infrastructure. With control over the sole verifier, the attacker was able to craft fake cross‑chain messages that appeared legitimate to the bridge’s relayers. These messages instructed the bridge to release large amounts of assets on the destination chain. Because the 1‑of‑1 configuration required no second opinion, the bridge executed the fraudulent transfers without additional checks, leading to the $292 million loss. The report emphasizes that had the 2‑of‑2 threshold remained, two independent verifiers would have had to collude, making the attack far more difficult.

What were the immediate consequences of the exploit, and how did LayerZero respond?

The immediate consequence was the loss of $292 million from the KelpDAO bridge, making it one of the largest bridge exploits in history. The stolen funds included a mix of ETH, USDC, and other tokens. Many users who had deposited assets into the bridge were left empty‑handed. LayerZero Labs quickly halted all bridge operations to prevent further losses and launched a forensic investigation with Mandiant, CrowdStrike, and zeroShadow. The team also notified law enforcement and partnered with centralized exchanges to freeze any illicit funds that could be traced. Additionally, LayerZero rolled out an emergency security patch for all DVN‑based bridges, enforcing a minimum 2‑of‑N threshold by default. The company has since advised all bridge operators to review their multi‑sig settings and implement time‑locked upgrades to prevent similar configuration changes.

What lessons can other cross‑chain bridges learn from this incident?

The KelpDAO exploit serves as a stark reminder that multi‑signature configurations are not just best practice but a critical layer of security for any decentralized bridge. First, never downgrade security thresholds without a rigorous change‑management process and multiple approvals. A single misconfiguration can undo robust protections. Second, bridge operators should implement on‑chain governance that requires a timelock or multisig vote before altering sensitive parameters like DVN thresholds. Third, regular security audits and penetration testing should be mandatory, especially after any upgrade. Fourth, incident response plans must include immediate ability to pause operations and freeze assets. Finally, the industry should move toward standardized, audited bridge architectures that prevent human error from introducing fatal vulnerabilities. The report underscores that security is not a one‑time setup but a continuous process of vigilance and adaptation.

How can users protect their assets when interacting with cross‑chain bridges?

For everyday users, the best defense is to research the security practices of any bridge before depositing funds. Look for bridges that publish transparency reports, undergo regular audits by reputable firms (like Mandiant or CrowdStrike), and use multi‑signature DVN configurations with a minimum 2‑of‑N threshold. Prefer bridges that implement timelocks for any configuration changes, giving users time to withdraw if an upgrade seems suspicious. Also, consider diversifying across multiple bridges to limit exposure if one is compromised. Monitor community channels and security bulletins for alerts. Finally, use hardware wallets and never share private keys. While users cannot control a bridge’s internal settings, they can vote with their feet—choosing only those bridges that prioritize robust, transparent security over speed or convenience.

Tags: