Skip to the stories
Microsoft 365 & SharePoint

Fake Recruiter Interviews Opened the Door to Binance's $6.3bn Breach

Binance says attackers spent months social-engineering its engineers through fake job interviews before a zero-day malware drop enabled the $6.3bn theft.

Fake Recruiter Interviews Opened the Door to Binance's $6.3bn Breach

The Binance breach is a reminder that attackers don't need to break your perimeter if they can persuade someone inside it to open the door. According to the exchange, the $6.3bn heist that began on 29 October was the end of a months-long campaign built on recruitment-themed social engineering, not a single exploit. For anyone running identity, endpoint or engineering access in a financial or crypto environment, the details are worth reading closely.

A phone screen is showing an app store with the details of the Binance app.

A North Korean group known as TraderTraitors is suspected of being behind the attack, which is widely regarded as the largest bank heist on record. Binance's further review found that the attackers combined several techniques over time, including fake recruitment calls and efforts to get engineers to download malicious software.

How the access was built

The timeline Binance has described starts in June. An engineer at the exchange downloaded a modified file. The company says this was discovered and fixed, but the incident still exposed sensitive information to the attackers. In other words, the first foothold was found and closed, and the attackers still walked away with data they could use later.

Net $3.6bn pulled by Binance users after multibillion-dollar heist

The more deliberate phase came in September. A second Binance engineer was contacted through a legitimate job-hunting platform by someone posing as a project director at an external company. That person arranged fake interviews, and the stated purpose was to build trust with the target before asking for anything.

At least 40 Irish credit unions targeted in significant cyberattack

Binance said the interviewer then shared a social media account with the engineer, including personal photos and videos posted there. A second round of interviews took place over social media. At the end of that process, the engineer was told to download a file containing malicious code onto his work computer.

$6.3bn hacked from Binance in ‘largest’ crypto attack

Here's the thing: that sequence doesn't need a sophisticated lure. It works because a recruitment process feels normal, the rapport feels earned, and a candidate assessment task feels like something a hiring manager would ask for. Teams in this position often treat a recruiter's request as low-risk because it sits outside the usual phishing pattern, and that's exactly the gap this approach exploits.

10 Irish cybersecurity experts you should know about

Where the zero-day got past the tools

Binance said the zero-day malware was not detected by its security protection tools at the time. It also said the attacker took steps to conceal the activity and remove traces of the compromise, which made detection and later investigation harder.

IOSCO: Financial regulators must prepare for quantum impact

That matters because it shifts the question from "did the endpoint protection fire?" to "what would we have seen if it hadn't?" Signature-based and reputation-based controls won't catch a novel payload. What teams usually need is behavioural telemetry on workstations, a way to spot unexpected child processes and persistence changes, and logs that are shipped off the host quickly enough that an attacker who cleans up locally can't erase the trail.

France’s officials suspect Iran over supermarket cyberattacks

The lateral movement phase is where the damage compounded. Nine days before the theft, Binance said, the attackers used the second engineer's compromised computer to take control of the devices of the first engineer and of another worker at the company. That step, from one workstation to two more staff devices, is the part I'd want every security team to map out in their own environment. Ask which engineering workstations can reach which admin paths, and whether a single compromised machine can pivot to people with elevated access.

What the exchange says it's doing about it

The fallout has been significant. Binance said some cryptocurrency coins dropped in value by thousands of dollars in the aftermath. Customers moved quickly to withdraw roughly $3.6bn from the exchange in recent days, and, as of the latest update on 3 November, Binance said it had frozen approximately $2.85bn in digital assets since the incident. The exchange has also said it has set aside "more than 100pc" of the funds required to cover the losses.

Binance also pointed to wider activity, saying its security team detected and prevented 57 suspected North Korea-linked cyberthreat cases throughout 2026. Its framing is that these attempts reflect a persistent threat from state-sponsored actors targeting the cryptocurrency industry. That's the company's characterisation, and the investigation into attribution rests on its own findings and on the suspicion that TraderTraitors is responsible.

What I'd check this week

My read is that the lesson here isn't "train staff not to click links." It's that recruitment channels, social media and personal accounts have become part of the attack surface for anyone with privileged access. A few practical questions follow from the Binance account:

First, do your engineers and admins have a clear rule for handling unsolicited recruiter contact and interview tasks, and does your HR process know to route any file-based assessment through a controlled path rather than a work laptop? Second, can an engineer's work device be reached from personal social media or messaging, and do you separate those contexts? Third, where do endpoint and identity logs go, and could a compromised workstation alter them before anyone notices? Fourth, if one engineering device is compromised, what other accounts and machines could it reach in a few steps?

Binance's account also shows that a first incident that gets contained can still leave sensitive information behind. If a June-style detection happened in your environment, would the response include a review of what the attacker could have seen, or would the ticket close once the file was removed? The exchange's own timeline suggests the answer should be no. Expect more of these multi-month, people-first campaigns to target the same kinds of staff, and plan your controls around the human path as well as the technical one.

Source: Suhasini Srinivasaragavan · www.siliconrepublic.com

Same flag