The scam targets online banking customers of several German banks. When the victim logs on to the online banking application, Tatanga uses a MitB webinject that alleges the bank is performing a security check on their computer and ability to receive a Transaction Authorization Number (TAN) on their mobile device.
In the background, Tatanga initiates a fraudulent money transfer to a mule account. It even checks the victim’s account balance, and will transfer funds from the account with the highest balance if there is more than one to choose from.
The victim is asked to enter the SMS-delivered TAN they receive from the bank into the fake web form, as a way to complete this security process. By entering the TAN in the injected HTML page the victim is in fact approving the fraudulent transaction originated by Tatanga against their account.
Even though the victim is presented with the fund transfer amount and the destination account information in the SMS message that contains the TAN, the injected HTML page claims that the process uses “experimental” data and that no money will leave their account.
Once the victim enters the TAN in the fake form and hits submit, the funds are transferred to the fraudster’s account. Meanwhile, Tatanga modifies the account balance reports in the online banking application to hide the fraudulent transaction.
“This is a very sophisticated and multi-faceted attack”, said Trusteer CTO Amit Klein. “By combining a MitB attack and social engineering, Tatanga is able to circumvent out-of-band authentication used by many banks. Then it goes one step further by hiding evidence of the fraudulent transaction from the victim using a post transaction attack mechanism.”
Fortunately, the text in the injected HTML page is littered with grammar and spelling mistakes and appears not to have been written by a German speaker. This may make it less effective. Clearly, grammar is easy for fraudsters to improve. The fact that they are blending multiple attack methods in a single fraud scam is not good news. However, they still need to compromise the endpoint with malware, which can be prevented.