The complete guide to Cross Site Scripting / XSS

What is Cross Site Scripting?
Cross-site scripting (XSS)is a type of computer security vulnerability typically found in web applications which allow code injection by malicious web users into the web pages viewed by other users. Cross-site scripting holes in general can be seen as vulnerabilities which allow attackers to bypass security mechanisms. By finding clever ways of injecting malicious scripts into web pages an attacker can gain elevated access privileges to sensitive page content, session cookies, and a variety of other objects.

There are three distinct types of XSS vulnerabilities:
non-persistent
persistent
and DOM-based (which can be either persistent or non-persistent).

Non-persistent cross-site scripting hole is also referred to as a reflected vulnerability, and is by far the most common type. These holes show up when data provided by a web client is used immediately by server-side scripts to generate a page of results for that user. A classic example of this is in site search engines: if one searches for a string which includes some HTML special characters, often the search string will be redisplayed on the result page to indicate what was searched for, or will at least include the search terms in the text box for easier editing. If any occurrence of the search terms is not HTML entity encoded, an XSS hole will result.

Persistent XSS vulnerability is also referred to as a stored or second-order vulnerability, and it allows the most powerful kinds of attacks. A type 2 XSS vulnerability exists when data provided to a web application by a user is first stored persistently on the server (in a database, file system, or other location), and later displayed to users in a web page without being encoded using HTML entities. A classic example of this is with online message boards, where users are allowed to post.

DOM-based XSS vulnerability, also referred to as local cross-site scripting, is based on the standard object model for representing HTML or XML called the Document Object Model or DOM for short. With DOM-based cross-site scripting vulnerabilities, the problem exists within a page's client-side script itself. For instance, if a piece of JavaScript accesses a URL request parameter and uses this information to write some HTML to its own page, and this information is not encoded using HTML entities, an XSS hole will likely be present, since this written data will be re-interpreted by browsers as HTML which could include additional client-side scripts.

Finding XSS Vulnerabilities
The most common used XSS injection test is:


alert("XSS")

When this example is injected into an input box or a URL parameter, it will either fire or it will fail. If the injection fails, it doesn't mean the site is secure, it just means you need to look deeper.

XSS Filter Evasion

Escaping From Strings
The first step is to view source on the Web page and see if you can find the injected string in the HTML.There are several places you may find it completely intact, yet hidden from the casual observer.The first is within an input parameter:

alert("XSS")

In this example we could alter our input to include two characters that allow the injected code to jump out of the single quotes:


'>alert("XSS")

Now our code renders because we have ended the input encapsulation and HTML tag before our vector, which allows it to fire. However, in this case, the extraneous single quote and closed angle bracket are displayed on the Web page.This can be suppressed if we update our vector into the following:


'>alert("XSS")alert("XSS")

As a result, the JavaScript code is injected with no visible indication of its existence.The tag does not render, because it is not valid.

Working Around Filtered Quotes
Let's use the same example above, but assume the Webmaster included code to put slashes in front of any single quotes or double quotes (i.e., add_slashes()). Our previous vector without the last part would now turn into:


\alert(\"XSS\")'>

There are several methods to try and work around this it all depends on the filtering in place. One method is to use Character Entities. Some characters are reserved in HTML. For example, you cannot use the greater than or less than signs within your text because the browser could mistake them for markup. If we want the browser to actually display these characters we must insert character entities in the HTML source.


" " " quotation mark, apl quote
& & & ampersand
< < < less-than sign
> > > greater-than sign

Using the code (") or (") in place of our quotes is one method to try and work around quote filtering. Example:


alert("XSS")
alert("XSS")
alert(&XSS&)

If no quotes of any kind are allowed you can use fromCharCode in JavaScript to create any XSS code you need. The fromCharCode() takes the specified Unicode values and returns a string. Example:


alert("XSS")
alert(String.fromCharCode(88,83,83))
\alert(String.fromCharCode(88,83,83))'>


You can use the For MySql char(ASCII,ASCII,...): calculator bellow to translate your code into CharCode.

Working Around Filtering
Some filters will filter out making it impossible for any of the above examples to work. However, there are many other ways to insert JavaScript into a Web page. Let's look at an example of an event handler:




The "onload" keyword inside HTML represents an event handler. It doesn't work with all HTML tags, but it is particularly effective inside BODY tags.That said, there are instances where this approach will fail, such as when the BODY onload event handler is previously overloaded higher on the page before your vector shows up. Another useful example is the onerror handler:




Because the image is poorly defined, the onerror event handler fires causing the JavaScript inside it to render, all without ever calling a tag.

Using IMG SRC
The two most commonly permitted HTML tags are




No quotes and no semicolon:




Filtering quotes and script:




Using CharCode to work around filtering quotes:




A simple attack vector, like the one above, can be even further obfuscated by transforming the entire string into the decimal equivalent of the ASCII characters:




Using the ASCII table you can decipher this example, and then use the same method of obfuscation to create your own injectable string. The same can be done for hexadecimal:




While the javascript: directive syntax inside images has been depreciated since IE 7.0, it still works in IE 6.0, Opera 9.0, Netscape 8.0 (when in the IE rendering engine, although it has also been depreciated as of 8.1)

Using Tab, New Line, and Carriage Return
Tab, new line and carriage return characters can also be used to trick XSS filters.




The example above uses a tab Minimum Sized Decimal to break up the word javascript intern breaking up the XSS and tricking the filter. The output above will look as follows:




ascript:alert('XSS');">


Horizontal Tab New line Carriage Return
URL %09 %10 %13
Minimal Sized Hex
Maximum Sized Hex
Minimum Sized Decimal
Maximum Sized Decimal

Using Null character
Another character that can cause problems for filters is the null character. This is one of the most obscure and powerful tools in any XSS arsenal. Take this example URL that can lead to a valid injection:


alert("XSS")

The null character (%00) stops the filters from recognizing the tag. This only works in IE 6.0, IE 7.0, and Netscape 8.0 in IE rendering engine mode.

Not filtering inside encapsulating pairs
Bypassing filtering that looks for open and closing pairs of encapsulation inside HTML tags and ignore the contents. Example:


alert('XSS')">

Technically, inside the IMG tag, the first two quotes should be considered encapsulation and should do nothing.The next quote should allow encapsulation and go to the next quote which is after the tag. Lastly, it should be closed by the trailing end angle bracket. But all major browsers, such as, IE, Firefox, Netscape, or Opera take this as malformed HTML and attempt to fix it. The output then looks like:


alert('xss')">

CSS Filter Evasion
HTML is a useful tool for injecting JavaScript, but not the only tool an even more complex sub-class of HTML is the style sheet or CSS. There are many different ways to inject XSS into style sheets, and even more ways to use them to inject JavaScript. . The simplest way to inject JavaScript into a CSS link tag is using the JavaScript directive.




However, IE has depreciated this as of 7.0, and it no longer works, you can still get it working in Opera and users who may still have IE 6.0 installed. Another way is to use the Using the above as an example, you can see how the expression tag allows the attacker to inject JavaScript without using the JavaScript directive or the tag.




Obscure Filters
Let's take an example where a developer has taken user input and insured that it contains no quotes, no angle brackets, and no JavaScript directives. Still, it is not safe, as we can inject something called a data directive in this case, we have base64 encoded the simple string alert('XSS').




The data directive allows us to inject entire documents inside a single string. The data directive works inside Firefox, Netscape in Gecko rendering engine mode, and Opera.

Using Double Quotes
If you need to use both double and single quotes you can use a grave accent to encapsulate the JavaScript string - this is also useful because lots of cross site scripting filters don't know about grave accents.




Escaping characters
Escaping quotes is sometimes usefull when there is an own written protection against XSS. This will allow you to escape the escape characters used by the XSS filter script.
It's worth mentioning that this will ONLY work if it's an own written (weak) defending script.




The result would be:




As you can see your own escape characters now filter out the escape characters used by the XSS protection.

Encoding

It is often assumed that if all angle brackets and quotes have been filtered that XSS is no longer possible. However XSS is reliant upon the browser, so as long as the browser can understand other encoding methods, you can run into situations where a browser will run commands without any of those characters.

A real world example of an XSS encoded vulnerability was found in Google search appliance by a hacker named Maluc. Maluc found that a normal Google search appliance query looked like:


[You must be registered and logged in to see this link.] ... anford&pro" target="_blank" rel="nofollow">[You must be registered and logged in to see this link.] xystylesheet=stanford&site=stanfordit&oe=UTF-8&q=hi

He noticed that according to this string (oe=UTF-Cool he could change the UTF code. He changed the UTF string from UTF-8 to UTF-7.

UTF-7 (7-bit Unicode Transformation Format) is a variable-length character encoding that was proposed for representing Unicode-encoded text using a stream of ASCII characters, for example for use in Internet e-mail messages. UTF-7 is generally not used as a native representation within applications as it is very awkward to process despite its size advantage over the combination of UTF-8 with either quoted-printable or base64.

Lets take for example:


alert("XSS")

And encode it using UTF-7:


+ADw-script+AD4-alert(+ACI-XSS+ACI-)+ADw-/script+AD4-

Now all + have to be changed to URL code in a GET strings for this to work. So the URL code for + is %2B now we have:


%2BADw-script%2BAD4-alert%281%29%2BADw-/script%2BAD4-

URL encoding is turning a string into a safe block of text for appending on the query string of a URL.To encode characters to append to a URL, you use a percentage symbol, followed by the two-digit hex number representing that character.

For example:


Original character Character Entity Reference
space %20
/ (forward slash) %2F
" (double quote) %22
? (question mark) %3F
+ %2B

With this Maluc came up with:


[You must be registered and logged in to see this link.] ... anford&pro" target="_blank" rel="nofollow">[You must be registered and logged in to see this link.] xystylesheet=stanford&site=stanfordit&oe=UTF-7&q=%2BADw-script%2BAD4-alert%281%29%2BADw-/script%2BAD4-x

And was able to successfully execute an XSS script.

Of course the effect of the XSS is only temporary and only affects the user who go to that URL, but this could easily provide an avenue for phishing. In this way, Google appliance has hurt Stanford University's security by being placed on the same domain.