A new report from the Arbor Security Engineering and Response Team (ASERT) sheds light on how the sophisticated banking malware known as Neverquest or Vawtrak is able to evade detection, by using encryption, anonymous routers, and even steganography.
"Neverquest is something we investigated because of its impact on the financial industry," said Kirk Soluk, manager of threat intelligence and responses at Arbor Networks. "Turns out, it's quite a sophisticated piece of malware. It does a lot of things to make it difficult for researchers to figure out how it works."
The malware has hit well over 100 of the world's biggest financial institutions in 25 countries.
It's installed on end user machines by specialized third-party malware installers, and then uses multiple avenues to get banking credentials, including looking at users' social media networks and email.
"The primary goal appears to be to steal credentials of customers that use these financial institutions," said Soluk.
"Once they have those credentials, the actors can actually log into the bank through the end user's PC, so it makes it very difficult for the financial institution to tell that something is wrong -- its not like they're logging in from some bizarre location. It will also try to manipulate transactions while its sitting on the customers' system. It's able to bypass a lot of the security techniques that people commonly think protect them."
But the malware doesn't just evade the banks' security, and the end user's anti-virus.
The malware writers have also layered in obfuscation techniques that makes it difficult for security researchers to figure out how the malware works -- and how to take it down.
"It's beyond just detection by, say, antivirus, which historically has had problems keeping up with mawalre on systems," said Soluk. "Just by them changing little bits here and there they can bypass antivirus -- if the user is even running up-to-date antivirus to begin with."
The malware uses encryption in the code itself to make it difficult to see what it is doing, he said.
It also uses encryption to communicate with its own command and control infrastructure.
And it also uses the TOR anonymity network and steganography -- a way of hiding information inside pictures.
By downloading seemingly harmless images, the malware can actually get updates to its command and control infrastructure, Soluk said.
Arbor has been peeling away the layers, starting with the first set of hard-coded command and control domains, and then moving to some of the Onion domains as well.
"We've identified 609 unique command and control locations," said Soluk, adding that the servers were mostly located in Western and Eastern Europe.
What particularly struck Soluk about this particular malware was the sheer amount of work that has gone into it.
"They're targeting between 100 and 200 different institutions, with hundreds of sites related to those institutions," he said. "And for each of those sites there are multiple web pages being targeted with web inject tools, and for each of those pages, you can have multiple inject points."
One reason that the authors have been able to dedicate these resources, he said, is because of the increased specialization in the malware industry.
Soluk added that Arbor's research confirmed an earlier report by Sophos that this particular malware is part of a malware-as-a-service operation.