DOM Based XSS

Definition

DOM Based XSS (or as it is called in some texts, “type-0 XSS”) is an XSS attack wherein the attack payload is executed as a result of modifying the DOM “environment” in the victim’s browser used by the original client side script, so that the client side code runs in an “unexpected” manner. That is, the page itself (the HTTP response that is) does not change, but the client side code contained in the page executes differently due to the malicious modifications that have occurred in the DOM environment.
This is in contrast to other XSS attacks (stored or reflected), wherein the attack payload is placed in the response page (due to a server side flaw).

Example

Suppose the following code is used to create a form to let the user choose his/her preferred language. A default language is also provided in the query string, as the parameter “default”.

…

Select your language:

<select><script>

document.write("<OPTION value=1>"+document.location.href.substring(document.location.href.indexOf("default=")+8)+"</OPTION>");

document.write("<OPTION value=2>English</OPTION>");

</script></select>
…

The page is invoked with a URL such as:

http://www.some.site/page.html?default=French

A DOM Based XSS attack against this page can be accomplished by sending the following URL to a victim:

http://www.some.site/page.html?default=<script>alert(document.cookie)</script>

When the victim clicks on this link, the browser sends a request for:

/page.html?default=<script>alert(document.cookie)</script>

to www.some.site. The server responds with the page containing the above Javascript code. The browser creates a DOM object for the page, in which the document.location object contains the string:

http://www.some.site/page.html?default=<script>alert(document.cookie)</script>

The original Javascript code in the page does not expect the default parameter to contain HTML markup, and as such it simply echoes it into the page (DOM) at runtime. The browser then renders the resulting page and executes the attacker’s script:

alert(document.cookie)

Note that the HTTP response sent from the server does not contain the attacker’s payload. This payload manifests itself at the client-side script at runtime, when a flawed script accesses the DOM variable document.location and assumes it is not malicious.

Advanced Techniques and Derivatives

In the example above, while the payload was not embedded by the server in the HTTP response, it still arrived at the server as part of an HTTP request, and thus the attack could be detected at the server side. The “DOM Based XSS” paper details a technique to avoid server side detection. It also describes several other possible locations for the payload, besides document.location.
The technique to avoid sending the payload to the server hinges on the fact that URI fragments (the part in the URI after the “#”) is not sent to the server by the browser. Thus, any client side code that references, say, document.location, may be vulnerable to an attack which uses fragments, and in such case the payload is never sent to the server. For example, the above DOM based XSS can be modified into:

http://www.some.site/page.html#default=<script>alert(document.cookie)</script>

which mounts the same attack without it being seen by the server (which will simply see a request for page.html without any URL parameters).
In December 2006, Stefano Di Paola and Giorgio Fedon described a universal XSS attack against the Acrobat PDF plugin ). This attack applied the fragment variant of DOM based XSS to PDF documents. The researchers discovered that a PDF document served to the browser, when rendered by the Acrobat plugin, may end up executing part of the fragment as Javascript. Since the Javascript is executed in the context (DOM) of the current site, all an attacker needed to exploit this flaw was to simply find a PDF link somewhere on the site for the XSS condition to be met. If the attacker then tricked a user into clicking on (or submitting) a link like:

http://www.some.site/somefile.pdf#somename=javascript:attackers_script_here

then a victim using an un-patched Acrobat reader would succumb to the attack. Adobe patched their reader after they were made aware of this flaw, but if not all users have downloaded the patch then those users are still vulnerable to this type of attack.

Extensions

Ory Segal gave an example (“Javascript flow manipulation” ) of how a target page can be framed and the frame’s parent (in the attacker’s control) can be devised in such manner that it affects the execution of the target page in a way desired by the attacker. The technique shows how DOM manipulation can be useful to modify the execution flow of scripts in the target page.

Testing Tools and Techniques

Minded Security has been doing some significant research into DOM based XSS. They are working on two projects to help with DOM Based XSS:
1. The DOMinator Tool - A commercial tool based on the Firefox browser with modified Spidermonkey Javascript engine that helps testers identify and verify DOM based XSS flaws
See: https://dominator.mindedsecurity.com/ (https://github.com/wisec/DOMinator for the open source part)
2. The DOM XSS Wiki - The start of a Knowledgebase for defining sources of attacker controlled inputs and sinks which could potentially introduce DOM Based XSS issues. Its very immature as of 11/17/2011. Please contribute to this wiki if you know of more dangerous sinks and/or safe alternatives!!
See: http://code.google.com/p/domxsswiki/
3. DOM Snitch - An experimental Chrome extension that enables developers and testers to identify insecure practices commonly found in client-side code. From Google.
See: http://code.google.com/p/domsnitch