OWASP Testing Guide Appendix D: Encoded Injection
This article is part of the OWASP Testing Guide v4 (the current status is:DRAFT).
OWASP Testing Guide v4 Table of Contents [DRAFT] At the moment the The entire OWASP Testing Guide v3 can be downloaded here.
Character Encoding is primarily used to represent characters, numbers and other symbols in a format that is suitable for a computer to understand, store, and render data. It is, in simple terms, the conversion of bytes into characters - characters belonging to different languages like English, Chinese, Greek or any other known language. A common and one of the early character encoding schemes is ASCII (American Standard Code for Information Interchange) that initially, used 7 bit coded characters. Today, the most common encoding scheme used is Unicode (UTF 8).
Character encoding has another use or rather misuse. It is being commonly used for encoding malicious injection strings in order to obfuscate and thus bypass input validation filters or take advantage of the browser’s functionality of rendering an encoding scheme.
Input Encoding – Filter Evasion
Web applications usually employ different types of input filtering mechanisms to limit the input that can be submitted by its users. If these input filters are not implemented sufficiently well, it is possible to slip a character or two through these filters. For instance, a / can be represented as 2F (hex) in ASCII, while the same character (/) is encoded as C0 AF in Unicode (2 byte sequence). Therefore, it is important for the input filtering control to be aware of the encoding scheme used. If the filter is found to be detecting UTF 8 encoded injections a different encoding scheme may be employed to bypass the filter.
In other words, an encoded injection works because even though an input filter might not recognize or filter an encoded attack, the browser correctly interprets it while rendering the web page.
Output Encoding – Server & Browser Consensus
Web browsers, in order to coherently display a web page, are required to be aware of the encoding scheme used. Ideally, this information should be provided to the browser through HTTP headers (“Content-Type”) as shown below:
Content-Type: text/html; charset=UTF-8
or through HTML META tag (“META HTTP-EQUIV”), as shown below:
<META http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">
It is through these character encoding declarations that the browser understands which set of characters to use when converting bytes to characters. Note: The content type mentioned in the HTTP header has precedence over the META tag declaration.
CERT describes it here as follows:
Many web pages leave the character encoding ("charset" parameter in HTTP) undefined. In earlier versions of HTML and HTTP, the character encoding was supposed to default to ISO-8859-1 if it wasn't defined. In fact, many browsers had a different default, so it was not possible to rely on the default being ISO-8859-1. HTML version 4 legitimizes this - if the character encoding isn't specified, any character encoding can be used.
If the web server doesn't specify which character encoding is in use, it can't tell which characters are special. Web pages with unspecified character encoding work most of the time because most character sets assign the same characters to byte values below 128. But which of the values above 128 are special? Some 16-bit character-encoding schemes have additional multi-byte representations for special characters such as "<". Some browsers recognize this alternative encoding and act on it. This is "correct" behavior, but it makes attacks using malicious scripts much harder to prevent. The server simply doesn't know which byte sequences represent the special characters
Therefore in the event of not receiving the character encoding information from the server, the browser either attempts to ‘guess’ the encoding scheme or reverts to a default scheme. In some cases, the user explicitly sets the default encoding in the browser to a different scheme. Any such mismatch in the encoding scheme used by the web page (server) and the browser may cause the browser to interpret the page in a manner that is unintended or unexpected.
All the scenarios given below form only a subset of the various ways obfuscation can be achieved in order to bypass input filters. Also, the success of encoded injections depends on the browser in use. For example, US-ASCII encoded injections were previously successful only in IE browser but not in Firefox. Therefore, it may be noted that encoded injections, to a large extent, are browser dependent.
Consider a basic input validation filter that protects against injection of single quote character. In this case the following injection would easily bypass this filter:
The above uses HTML Entities to construct the injection string. HTML Entities encoding is used to display characters that have a special meaning in HTML. For instance, ‘>’ works as a closing bracket for a HTML tag. In order to actually display this character on the web page HTML character entities should be inserted in the page source. The injections mentioned above are one way of encoding. There are numerous other ways in which a string can be encoded (obfuscated) in order to bypass the above filter.
A variation of the above string is given below. Can be used in case ‘%’ is being filtered:
There are other encoding schemes like Base64 and Octal as well that may be used for obfuscation. Although, every encoding scheme may not work every time, a bit of trial and error coupled with intelligent manipulations would definitely reveal the loophole in a weakly built input validation filter.
UTF-7 encoding of <SCRIPT>alert(‘XSS’);</SCRIPT> is as below
For the above script to work, the browser has to interpret the web page as encoded in UTF-7.
Variable-width encoding is another type of character encoding scheme that uses codes of varying lengths to encode characters. Multi-Byte Encoding is a type of variable-width encoding that uses varying number of bytes to represent a character. Multibyte encoding is primarily used to encode characters that belong to a large character set e.g. Chinese, Japanese and Korean.
Multibyte encoding has been used in the past to bypass standard input validation functions and carry out cross site scripting and SQL injection attacks.