North Korean hackers drained about $6.34 billion from Binance in roughly 20 minutes on October 29, and it began with a fake job interview.
According to Binance’s November 3 report, attackers infected an engineer’s computer, used an internal AI tool to reach two more employees, and hijacked the exchange’s wallet-signing service. Security experts say four basic controls could have stopped them.
Inside the Binance hack: how attackers reached the wallets
Unlike conventional cryptocurrency thefts involving stolen private keys or direct attacks on smart contracts, the Binance incident reportedly involved a multi-stage compromise of the exchange’s internal infrastructure.
According to Binance’s November 3 report, the attackers initially compromised the computer of an engineer identified as Engineer B through a malicious recruitment exercise. The operation reportedly involved fake interview-related activities and malicious software downloads.
The attackers subsequently used an internal artificial intelligence tool to gain access to the computers of two additional employees, including a wallet DevOps engineer and a software engineer.
The compromised wallet DevOps engineer had a container deployment permission token stored on his computer. This provided an opportunity to access systems involved in deploying software and interacting with wallet infrastructure.
Binance said the attackers used stolen credentials to deploy a container and execute a malicious script that reached its wallet-signing operations.
On October 29, the attackers initiated unauthorised transfers across multiple blockchain networks. On Ethereum, they exploited the EIP-7702 authorisation mechanism to grant an attacker-controlled smart contract execution permissions over a Binance hot wallet.
The attack affected wallets across ten blockchain networks, with most of the theft occurring within approximately 20 minutes.
Blockchain intelligence firm TRM Labs attributed the operation to North Korea’s TraderTraitor group. The FBI also identified North Korea as responsible for the incident.
Where the Binance hack could have been stopped
The first opportunity to prevent the theft was at the employee-device level.
A compromised computer should not automatically provide access to critical infrastructure, even when its user possesses legitimate organisational permissions.
Stronger endpoint detection, application restrictions, device isolation and continuous monitoring could have helped identify malicious software before it spread across the organisation.
The second weakness involved access permissions. The attackers reportedly obtained credentials that allowed them to deploy containers connected to wallet operations.
A more restrictive access-control system could have limited what those credentials were permitted to do. Short-lived authentication tokens, hardware-backed credentials, just-in-time access and mandatory independent approval for production deployments would have reduced the consequences of compromised employee devices.
The third opportunity involved Binance’s internal AI tooling. The investigation indicates that the attackers used an internal AI tool as part of their movement between employee computers.
AI systems connected to internal company infrastructure should operate within strict permission boundaries. They should not automatically inherit access to sensitive employee environments, deployment credentials or wallet-management systems.
Separating AI environments from privileged infrastructure, restricting access to internal resources and monitoring unusual tool interactions could have made this stage of the operation more difficult.
How the Binance hack could have been prevented at wallet level
Perhaps the most important lesson concerns transaction authorisation.
Binance’s preliminary findings indicate that the attackers reached the internal wallet-signing service and initiated transactions outside the normal withdrawal process.
A properly isolated signing system should independently verify every transaction before releasing an authorisation. Employee credentials or access to a deployment environment should not, on their own, be sufficient to initiate large withdrawals.
For example, a transaction policy engine could reject transfers that exceed predefined limits, involve unfamiliar smart contracts or originate from unexpected deployment environments.
Additional safeguards could include transaction allowlists, independent approvals, withdrawal delays and separate authorisation systems for different blockchain networks.
Real-time monitoring could also have detected unusual transfers across multiple networks and automatically suspended wallet operations before the losses escalated.
Binance’s incident demonstrates why storing assets in cold wallets or using sophisticated cryptographic authorisation mechanisms cannot replace security throughout the entire operational environment.
What the Binance hack means for crypto security
Binance acknowledged the seriousness of the incident in its November 3 statement.
“This should not have happened, and we apologise for the concern, disruption and uncertainty it caused our users, partners and the wider industry,” the exchange said.
The company subsequently reported that approximately $2.85 billion in assets had been frozen or intercepted, subject to reconciliation and market price movements. It also said it had replenished 98% of the affected assets as of October 31.
Binance announced that it had rebuilt affected systems, replaced compromised devices, revoked exposed credentials and strengthened wallet-service controls and behavioural anomaly detection.
The exchange also committed to independent cybersecurity reviews.
The FBI’s attribution of the operation to North Korea highlights the growing sophistication of state-linked cybercriminal operations targeting digital asset infrastructure.