Testing for Stack Traces (OWASP-IG-XXX)

Brief Summary
Stack traces are not vulnerabilities by themselves, but they often reveal information that is interesting to an attacker. Attackers attempt to generate these stack traces by tampering with the input to the web application with malformed HTTP requests and other input data.

Description of the Issue
If the application responds with stack traces that are not managed it could reveals potentially useful information for an attacker. This information could then be used, by the attacker, to bring successive attacks. Provide debugging information as a result of operations that generate errors is considered a bad practice due to the fact that relative paths to the point where the application is installed or how objects are referenced internally are shown to the user which can draw benefit for a successive attack.

Black Box testing and example
There are a variety of techniques that will cause exception messages to be sent in an HTTP response. Note that in most cases this will be an HTML page, but exceptions can be sent as part of SOAP or REST responses too.

Some tools, such as OWASP ZAP and Burp proxy will automatically detect these exceptions in the response stream as you are doing other penetration and testing work.

Gray Box testing and example
Search the code for the calls that cause an exception to be rendered to a String or output stream. For example, in Java this might be code in a JSP that looks like:

&lt;% e.printStackTrace( new PrintWriter( out ) ) %&gt;

In some cases, the stack trace will be specifically formatted into HTML, so be careful of accesses to stack trace elements.

Search the configuration to verify error handling configuration and the use of default error pages. For example, in Java this configuration can be found in web.xml.