Use of Implicit Intent for Sensitive Communication
The Android application uses an implicit intent for transmitting sensitive data to other applications.
Since an implicit intent does not specify a particular application to receive the data, any application can process the intent by using an Intent Filter for that intent. This can allow untrusted applications to obtain sensitive data. There are two variations on the standard broadcast intent, ordered and sticky.
Ordered broadcast intents are delivered to a series of registered receivers in order of priority as declared by the Receivers. A malicious receiver can give itself a high priority and cause a denial of service by stopping the broadcast from propagating further down the chain. There is also the possibility of malicious data modification, as a receiver may also alter the data within the Intent before passing it on to the next receiver. The downstream components have no way of asserting that the data has not been altered earlier in the chain.
Sticky broadcast intents remain accessible after the initial broadcast. An old sticky intent will be broadcast again to any new receivers that register for it in the future, greatly increasing the chances of information exposure over time. Also, sticky broadcasts cannot be protected by permissions that may apply to other kinds of intents.
In addition, any broadcast intent may include a URI that references data that the receiving component does not normally have the privileges to access. The sender of the intent can include special privileges that grant the receiver read or write access to the specific URI included in the intent. A malicious receiver that intercepts this intent will also gain those privileges and be able to read or write the resource at the specified URI.
The following examples help to illustrate the nature of this weakness and describe methods or techniques which can be used to mitigate the risk.
Note that the examples here are by no means exhaustive and any given weakness may have many subtle varieties, each of which may require different detection methods or runtime controls.
This application wants to create a user account in several trusted applications using one broadcast intent:
This application assumes only the trusted applications will be listening for the action. A malicious application can register for this action and intercept the user's login information, as below:
When a broadcast contains sensitive information, create an allowlist of applications that can receive the action using the application's manifest file, or programmatically send the intent to each individual intended receiver.
This application interfaces with a web service that requires a separate user login. It creates a sticky intent, so that future trusted applications that also use the web service will know who the current user is:
Sticky broadcasts can be read by any application at any time, and so should never contain sensitive information such as a username.
This application is sending an ordered broadcast, asking other applications to open a URL:
Any application in the broadcast chain may alter the data within the intent. This malicious application is altering the URL to point to an attack site:
The final receiving application will then open the attack URL. Where possible, send intents to specific trusted applications instead of using a broadcast chain.
This application sends a special intent with a flag that allows the receiving application to read a data file for backup purposes.
Any malicious application can register to receive this intent. Because of the FLAG_GRANT_READ_URI_PERMISSION included with the intent, the malicious receiver code can read the user's data.
This view (slice) covers all the elements in CWE.
CWE identifiers in this view are weaknesses that do not have associated Software Fault Patterns (SFPs), as covered by the CWE-888 view. As such, they represent gaps in...
CWE entries in this view (slice) are often seen in mobile applications.