Less than three weeks after North Korea-linked hackers used social engineering to hit crypto trading firm Drift, hackers tied to the nation appear to have pulled off another major exploit with Kelp.
The attack on Kelp, a restaking protocol tied into LayerZero’s cross-chain infrastructure, suggests an evolution in how North Korea-linked hackers operate, not just looking for bugs or stolen credentials, but exploiting the basic assumptions built into decentralized systems.
Taken together, the two incidents point to something more organized than a string of one-off hacks, as North Korea continues to escalate its efforts to hijack funds from the crypto sector.
“This is not a series of incidents; it is a cadence,” said Alexander Urbelis, chief information security officer and general counsel at ENS Labs. “You cannot patch your way out of a procurement schedule.”
More than $500 million was siphoned across the Drift and Kelp exploits in just over two weeks.
How Kelp was breached
At its core, the Kelp exploit did not involve breaking encryption or cracking keys. The system actually worked the way it was designed to. Rather, attackers manipulated the data feeding into the system and forced it to rely on those compromised inputs, causing it to approve transactions that never actually occurred.
“The security failure is simple: a signed lie is still a lie,” Urbelis said. “Signatures guarantee authorship; they do not guarantee truth.”
In simpler terms, the system checked who sent the message, not whether the message itself was correct. For security experts, that makes this less about a clever new hack and more about exploiting how the system was set up.
“This attack wasn’t about breaking cryptography,” said David Schwed, COO of blockchain security firm SVRN. “It was about exploiting how the system was set up.”
One key issue was a configuration choice. Kelp relied on a single verifier, essentially one checker, to approve cross-chain messages. That is because it’s faster and simpler to set up, but it removes a critical safety layer.
LayerZero has since recommended using multiple independent verifiers to approve transactions in the fallout, similar to requiring multiple signatures on a bank transfer. Some in the ecosystem have pushed back on that framing, saying that LayerZero’s default setup was to have a single verifier.
“If you’ve identified a configuration as unsafe, don’t ship it as an option,” Schwed said. “Security that depends on everyone reading the docs and getting it right is not realistic.”
The fallout has not stayed limited to Kelp. Like many DeFi systems, its assets are used across multiple platforms, meaning problems can spread.
“These assets are a chain of IOUs,” Schwed said. “And the chain is only as strong as the controls on each link.”
When one link breaks, others are affected. In this case, lending platforms like Aave that accepted the impacted assets as collateral are now dealing with losses, turning a single exploit into a wider stress event.
Decentralization marketing
The attack also exposes a gap between how decentralization is marketed and how it actually works.
“A single verifier is not decentralized,” Schwed said. “It’s a centralized decentralized verifier.”
Urbelis puts it more broadly.
“Decentralization is not a property a system has. It is a series of choices,” he said. “And the stack is only as strong as its most centralized layer.”
In practice, that means even systems that appear decentralized can have weak points, especially in the less visible layers like data providers or infrastructure. Those are increasingly where attackers are focusing.
That shift may explain Lazarus’ recent targeting.
The group has begun zeroing in on cross-chain and restaking infrastructure, Urbelis said, the parts of crypto that move assets between systems or allow them to be reused.
These layers are critical but complex, often sitting underneath more visible applications. They also tend to hold large amounts of value, making them attractive targets.
If earlier waves of crypto hacks focused on exchanges or obvious code flaws, recent activity suggests a move toward what could be called the industry’s plumbing, the systems that connect everything together, but are harder to monitor and easier to misconfigure.
As Lazarus continues to adapt, the biggest risk may not be unknown vulnerabilities, but known ones that are not fully addressed.
The Kelp exploit did not introduce a new kind of weakness. It showed how exposed the ecosystem remains to familiar ones, especially when security is treated as a recommendation rather than a requirement.
And as attackers move faster, that gap is becoming both easier to exploit and far more expensive to ignore.