German university researchers have found hundreds of popular Android apps in the Google Play market that leave 10s of millions of users vulnerable to attackers looking to steal banking credentials, credit card numbers and other personal information.
The problem is in the way the tablet and smartphone apps implement the security protocol used in communicating with users' Web browsers, the researchers said. An analysis of thousands of free apps found nearly 8% vulnerable to man-in-the-middle attacks as a result of unsound use of secure socket layer (SSL).
In general, mobile apps use transport layer security (TLS), which includes the SSL protocol, for transmitting and receiving sensitive data while communicating with a Web server. The researchers claim that flaws in the implementation make it possible for an attacker to intercept and control the data traffic.
[See also from Antone Gonsalves: FBI warns commercial spyware has made jump to Android]
During the analysis, researchers were able to intercept from the apps a variety of user information, such as credit card numbers, bank account information, PayPal credentials and social network credentials.
Google did not respond to a request for comment.
The full research paper was not immediately available, but the authors did post an overview of the findings. In analyzing the use of SSL in 13,500 popular free apps on Google Play, the researchers found 1,074 vulnerable to MITM attacks.
The researchers, who worked in teams from the Leibniz University in Hanover and Philipps University of Hamburg, used a homegrown proof-of-concept tool called MalloDroid, which was designed to identify exploitable SSL bugs, Threatpost, the Kaspersky Lab blog, reported Friday. The apps analyzed with the tool represented 17% of the apps that contain HTTPS URLs, which indicate that they use SSL.
The researchers manually audited 100 apps and found 41 vulnerable to MITM attacks because of SSL misuse. The cumulative install base of all the vulnerable apps ranged between 39.5 million and 185 million users, based on information the researchers gathered from Google Play.
"The actual number is likely to be larger, since alternative app markets for Android also contribute to the install base," the researchers said.
From the 41 apps analyzed manually, the researchers were able to capture credentials for American Express, Diners Club, Paypal, bank accounts, Facebook, Twitter, Google, Yahoo, Microsoft Live ID, Box, WordPress, remote control servers, arbitrary email accounts and IBM Sametime. In addition, the researchers were able to disable anti-virus apps and remotely inject and execute code.
The research findings were not a surprise to Jeremiah Grossman, founder and chief technology officer of WhiteHat Security.
"The implementation problems with SSL was, or is, true with a good number of Websites as well," Grossman said in an email. "As the mobile landscape is immature, one might expect this to be the case."
Chester Wisniewski, senior security adviser for Sophos, said the problem often stems from an application developer who does not know how to properly implement SSL. Also, careless developers skip implementing SSL in the beta version of an app and never turn it on when the app becomes generally available.
"One of the problems with SSL is it's very fragile," Wisniewski said. "If you break any one piece of how it works because it's inconvenient, and disable (the component) or turn it off, then the whole thing is useless."
For companies, the research points to the importance of using virtual private networks when employees are accessing corporate email and data using a mobile device, Wisniewski said. "That way, it would be very difficult for somebody to man-in-the-middle it."
For consumers, Wisniewski recommended using the Chrome browser in an Android device or Apple Safari in the iPhone for sensitive transactions, when possible. The browsers have more reliable SSL implementations then apps.
In addition, people should avoid using public Wi-Fi networks for making purchases with a credit or debit card or accessing online banking, since most MITM attacks occur on such networks.
The latest research adds to the risks associated with Android devices. Because apps can be sold and distributed by anyone, a significant number have been found to be malware capable of stealing data, sending texts to paid services and distributing spam.
Despite the risks, wireless carriers and Android device makers continue to do a poor job at patching the software, recent studies show.
Read more about wireless/mobile security in CSOonline's Wireless/Mobile Security section.