Security Flaw Left Major Banking Apps Vulnerable to MiTM Attacks Over SSL

A team of security researchers has discovered a critical implementation flaw in major mobile banking applications that left banking credentials of millions of users vulnerable to hackers.

The vulnerability was discovered by researchers of the Security and Privacy Group at the University of Birmingham, who tested hundreds of different banking apps—both iOS and Android—and found that several of them were affected by a common issue, leaving their users vulnerable to man-in-the-middle attacks.

The affected banking apps include HSBC, NatWest, Co-op, Santander, and Allied Irish bank, which have now been updated after researchers reported them of the issue.

According to a research paper [PDF] published by researchers, vulnerable applications could have allowed an attacker, connected to the same network as the victim, to intercept SSL connection and retrieve the user's banking credentials, like usernames and passwords/pincodes—even if the apps are using SSL pinning feature.

SSL pinning is a security feature that prevents man-in-the-middle (MITM) attacks by enabling an additional layer of trust between the listed hosts and devices.

When implemented, SSL pinning helps to neutralize network-based attacks wherein attackers could attempt to use valid certificates issued by rogue certification authorities.

"If a single CA acted maliciously or were compromised, which has happened before, valid certificates for any domain could be generated allowing an attacker to Man-in-the-Middle all apps trusting that CA certificate," the researchers wrote in their paper.

However, there are two key parts to verify an SSL connection—the first (authentication) is to verify whether the certificate is from a trusted source and the second (authorization) is to make sure the server you are connecting to presents the right certificate.

Researchers found that due to lack of hostname verification, several banking applications were not checking if they connected to a trusted source.

Verifying a hostname ensures the hostname in the URL to which the banking app connects matches the hostname in the digital certificate that the server sends back as part of the SSL connection.

"TLS misconfiguration vulnerabilities are clearly common; however none of the existing frameworks will detect that a client pins a root or intermediate certificate, but fails to check the hostname in the leaf," the paper reads.

Besides this issue, the researchers also detailed an "in-app phishing attack" affecting Santander and Allied Irish Banks, which could have allowed attackers to hijack part of the victim's screen while the app was running and use it to phish for the victim's login credentials.

To test this vulnerability in hundreds of banking apps quickly and without requiring to purchase certificates, researchers created a new automated tool, dubbed Spinner.
Spinner leverages Censys IoT search engine for finding certificate chains for alternate hosts that only differ in the leaf certificate.

"Given the certificate for a target domain, the tool queries for certificate chains for alternate hosts that only differ in the leaf certificate. The tool then redirects the traffic from the app under test to a website which has a certificate signed by the same CA certificate, but of course a different hostname (Common Name)," the researchers explain.

"If the connection fails during the establishment phase then we know the app detected the wrong hostname. Whereas, if the connection is established and encrypted application data is transferred by the client before the connection fails then we know the app has accepted the hostname and is vulnerable."

The trio, Chris McMahon Stone, Tom Chothia, and Flavio D. Garcia, worked with the National Cyber Security Centre (NCSC) to notify all affected banks, which then resolved the issues before they publicly disclosed their research this week.