Researchers Link Magecart Group 4 to Cobalt Group
Their findings demonstrate how Group 4 is likely conducting server-side skimming in addition to client-side activity.
Security researchers have discovered a link between Magecart Group 4 and Cobalt Group, a well-known, financially motivated group in operation since 2015. Findings indicate Group 4 is not only conducting client-side skimming but was, and likely still is, doing the same server-side.
Magecart is an umbrella term for at least seven cybercriminal groups responsible for installing skimmers onto e-commerce websites with the goal of lifting payment data. The threat quickly gained notoriety as groups planted skimmers on websites including British Airways and Ticketmaster.
Researchers across cybersecurity have been ramping up efforts to connect Magecart operators with known attack groups. This initiative started with RiskIQ and Flashpoint's Inside Magecart research; more recently, IBM publicly connected Group 6 with FIN6. Now researchers with Malwarebytes and security firm HYAS have found patterns linking Group 4 with Cobalt Group.
Group 4 is one of the more advanced groups. Its operators use sophisticated techniques to blend into normal traffic — for example, registering domain names seemingly connected with analytic providers or advertisers. The group also seems to have a history in banking malware, the same area of expertise as Cobalt Group, otherwise known as FIN7 and Carbanak Group.
"We knew that Group 4 was different from other groups based on their skill level," says Jerome Segura, head of threat intelligence at Malwarebytes. "We knew it was not just an average threat actor that decided one day to do some skimming."
Given its expertise and track record with banking malware, the researchers' attention turned to APT groups to find a connection. They reached out to HYAS, which reported it had also linked Magecart to an APT group.
Malwarebytes had been tracking the different Magecart groups and trying to find a trail, Segura continues. Researchers looked for pieces of infrastructure used by Magecart groups, as well as connections between domain registrations and IP locations. They used indicators of compromise, domain registration, and code from TTPs (tactics, techniques, and procedures) to conclude that Cobalt Group may have shifted to Web skimming.
Both the client-side and server-side skimmer domains were registered to a protonmail address that RiskIQ researchers had linked to Group 4. Researchers checked their exfiltration gates and connected these domains to other registrant emails. They noticed the email addresses used to register Magecart domains for Group 4 followed a certain pattern of [firstname] [initial] [lastname] – the same pattern Cobalt Group had recently adopted with protonmail accounts. In addition to the same email service, Cobalt Group was also using the same privacy protection.
"Given the use of privacy services for all the domains in question, it is highly unlikely that this naming convention would be known to any other actor besides those who registered both the Cobalt Group and Magecart infrastructure," researchers said in a writeup of their findings. What's more, 10 of the seemingly separate accounts only reused two of the same IP addresses.
Server-Side Skimming
While checking infrastructure related to Group 4, Malwarebytes identified a PHP script they think was mistakenly served as JavaScript. It's the type of source code you're not supposed to see unless you have access to the server, and it's a script that exclusively skims server-side.
"It's invisible to any scanner or crawler because everything happens on the actual compromised server," Segura explains. Magecart reports typically focus on browser-side skimmers; server-side skimmers aren't often covered because many companies don't have visibility into the back-end servers of compromised websites. "These are much harder to detect," he points out.
Client-side skimming is easier for attackers to do and for defenders to detect. The malicious JavaScript skimmer usually loads at one of two points: when a shopper heads to checkout or when they transmit their banking details. At either of these points, security software can jump in and stop the malicious script from loading. Most tools can block domain names, IP addresses, or malicious code because the activity is always happening within the browser, Segura explains.
In server-side skimming, there is no JavaScript loaded into the broader; data exfiltration happens entirely in the server. For defenders who want to block malicious activity on a machine, there is nothing to see. If the server has been compromised, it's already too late.
As far as Web skimming goes, most of the discussion revolves around client-side activity because the skimmers are available for purchase online and are easier to customize. Attackers can track their activities and automate and scale. It's "fairly easy" to rent a skimmer kit, buy some exploits to compromise sites, and start skimming, which explains the recent increase of newcomers in the space, Segura says. Server-side skimming, in comparison, involves more effort to customize and maintain each skimmer for every e-commerce website targeted, he adds.
Over time, as this activity continues to mature, Segura says he anticipates we'll see more players and a greater spectrum of sophistication as advanced attackers specialize in areas of expertise. It's already happening, he says. Some groups are going after high-value targets, foregoing smaller sites in favor of plug-in providers, analytics providers, and other parts of the supply chain.
Related Content:
Check out The Edge, Dark Reading's new section for features, threat data, and in-depth perspectives. Today's top story: "How the City of Angels Is Tackling Cyber Devilry."
About the Author
You May Also Like
A Cyber Pros' Guide to Navigating Emerging Privacy Regulation
Dec 10, 2024Identifying the Cybersecurity Metrics that Actually Matter
Dec 11, 2024The Current State of AI Adoption in Cybersecurity, Including its Opportunities
Dec 12, 2024Cybersecurity Day: How to Automate Security Analytics with AI and ML
Dec 17, 2024The Dirt on ROT Data
Dec 18, 2024