MySQL Injection: Wide Byte Injection (GBK)

Wide byte injection is a specific type of SQL injection attack that targets applications using multi-byte character sets, like GBK or SJIS. The term "wide byte" refers to character encodings where one character can be represented by more than one byte. This type of injection is particularly relevant when the application and the database interpret multi-byte sequences differently. The `SET NAMES gbk` query can be exploited in a charset-based SQL injection attack. When the character set is set to GBK, certain multibyte characters can be used to bypass the escaping mechanism and inject malicious SQL code. Several characters can be used to trigger the injection. * `%bf%27`: This is a URL-encoded representation of the byte sequence `0xbf27`. In the GBK character set, `0xbf27` decodes to a valid multibyte character followed by a single quote ('). When MySQL encounters this sequence, it interprets it as a single valid GBK character followed by a single quote, effectively ending the string. * `%bf%5c`: Represents the byte sequence `0xbf5c`. In GBK, this decodes to a valid multi-byte character followed by a backslash (`\`). This can be used to escape the next character in the sequence. * `%a1%27`: Represents the byte sequence `0xa127`. In GBK, this decodes to a valid multi-byte character followed by a single quote (`'`). A lot of payloads can be created such as:

%A8%27 OR 1=1;--
%8C%A8%27 OR 1=1--
%bf' OR 1=1 -- --

Here is a PHP example using GBK encoding and filtering the user input to escape backslash, single and double quote.

function check_addslashes($string)
{
    $string = preg_replace('/'. preg_quote('\\') .'/', "\\\\\\", $string);          //escape any backslash
    $string = preg_replace('/\'/i', '\\\'', $string);                               //escape single quote with a backslash
    $string = preg_replace('/\"/', "\\\"", $string);                                //escape double quote with a backslash
      
    return $string;
}

$id=check_addslashes($_GET['id']);
mysql_query("SET NAMES gbk");
$sql="SELECT * FROM users WHERE id='$id' LIMIT 0,1";
print_r(mysql_error());

Here's a breakdown of how the wide byte injection works: For instance, if the input is `?id=1'`, PHP will add a backslash, resulting in the SQL query: `SELECT * FROM users WHERE id='1\'' LIMIT 0,1`. However, when the sequence `%df` is introduced before the single quote, as in `?id=1%df'`, PHP still adds the backslash. This results in the SQL query: `SELECT * FROM users WHERE id='1%df\'' LIMIT 0,1`. In the GBK character set, the sequence `%df%5c` translates to the character `連`. So, the SQL query becomes: `SELECT * FROM users WHERE id='1連'' LIMIT 0,1`. Here, the wide byte character `連` effectively "eating" the added escape character, allowing for SQL injection. Therefore, by using the payload `?id=1%df' and 1=1 --+`, after PHP adds the backslash, the SQL query transforms into: `SELECT * FROM users WHERE id='1連' and 1=1 --+' LIMIT 0,1`. This altered query can be successfully injected, bypassing the intended SQL logic.

Source: SQL Injection/MySQL Injection.md in PayloadsAllTheThings. From PayloadsAllTheThings by Swissky and contributors. MIT License, Copyright (c) 2019 Swissky.