M8: Security Decisions Via Untrusted Inputs

Threat Agents

Application Specific

Threat Agents include entities that can pass untrusted inputs to the sensitive method calls. Examples of such entities include, but are not limited to, users, malware and vulnerable apps.

Attack Vectors

Exploitability EASY

An attacker with access to app can intercept intermediate calls and manipulate results via parameter tampering.

Security Weakness

Prevalence COMMON
Detectability EASY

Developers generally use hidden fields and values or any hidden functionality to distinguish higher level users from lower level users. An attacker can intercept the calls (IPC or web service calls) and temper with such sensitive parameters. Weak implementation of such functionalities leads to improper behavior of an app and even granting higher level permissions to an attacker.This can easily be exploited through hooking functionality.

Technical Impacts

Impact SEVERE

This vulnerability may lead to privilege escalation providing access of higher authorities and functionalities to an attacker. It can even bypass security mechanisms implemented by the app leading to loss of confidentiality and integrity.

Business Impacts

Application / Business Specific

This vulnerability leads to loss of reputation. At the same time, impacting and harming the integrity and confidentiality.

Am I Vulnerable To ‘Security Decisions Via Untrusted Inputs’?

Your mobile application can accept data from all kinds of sources. In most cases this will be an Inter Process Communication (IPC) mechanism. In general try and adhere to the following IPC design patterns:

  • If there is a business requirement for IPC communication, the mobile application should restrict access to a white-list of trusted applications
  • Sensitive actions which are triggered through IPC entry points should require user interaction before performing the action
  • All input received from IPC entry points must undergo stringent input validation in order to prevent input driven attacks
  • Do not pass any sensitive information through IPC mechanisms, as it may be susceptible to being read by third party applications under certain scenarios

How Do I Prevent ‘Security Decisions Via Untrusted Inputs’?

iOS Specific Examples:

  • Do not use the deprecated handleOpenURL method to handle URL Scheme calls. This method does not contain an argument containing the BundleID of the source application.
    • Instead use the openURL:sourceApplication:annotation method and validation the sourceApplication argument against a white-list of trusted applications
  • Do not use the iOS Pasteboard for IPC communications, as it is susceptible to being set or read by all third party apps on the device.

Android Specific Examples

  • None

Example Attack Scenarios

None

References