Using the vBulletin Input Cleaner
 

Note: This article assumes that you are familier with PHP, and will introduce you to input filtering using vBulletin


Introduction

Most scripts will require data from a user at some point. When using this data, you should never assume that it is "clean" data. With XSS (Cross-Site Scripting) and SQL exploits being identified in scripts on a daily basis, you should do everything you can to ensure that all data coming from the user has been cleaned ("sanatized").

vBulletin provides us with the vB_Input_Cleaner class to do just this.
The vBulletin Input Cleaner class is setup when the page loads, and can be accessed as $vbulletin->input,

Data Types

When you accept data from the user, you should know what type of data you are expecting to receive. the vBulletin Input Cleaner allows the following types of data to be cleaned:

• TYPE_NOCLEAN
  Will not be cleaned
• TYPE_BOOL
  Will check it is either true or false
• TYPE_INT
  Will check that it is an integer
• TYPE_UINT
  Will check that it is an unsigned integer
• TYPE_NUM
  Will check that it is a number
• TYPE_UNUM
  Will check that it is an unsigned number
• TYPE_UNIXTIME
  Will check that it is a unix-style timestamp (unsigned int)
• TYPE_STR
  Will check that it is a string, and runs trim() on it
• TYPE_NOTRIM
  Will check that it is a string and will not run trim() on it
• TYPE_NOHTML
  WIll check that it is a string and run htmlspecialchars_uni() and trim() on it
• TYPE_ARRAY
  WIll check that it is an array
• TYPE_FILE
  Will check that it is a file (ie, uploaded by the user)

You can also clean arrays of these types by using TYPE_ARRAY_<type>. For example, if you had an array of numbers, you could use TYPE_ARRAY_INT, or TYPE_ARRAY_NUM.

Cleaning Functions

The input cleaner class provides a number of useful functions that we can use to clean our data, depending on what data you wish to clean.

Cleaning Superglobal Arrays

By Superglobal, I mean $_POST, $_GET, $_REQUEST and so on. These arrays are created automaticly by PHP and contain the user-sent input. They are referenced in the vBulletin Input Cleaner by nice short single letter names. These are:

• p - $_POST
• g - $_GET
• r - $_REQUEST
• s - $_SERVER
• e - $_ENV
• c - $_COOKIE
• f - $_FILES

The vBulletin Input Cleaner class provides the clean_array_gpc() function which allows us to clean data in these Superglobal arrays in one hit, without having to clean every individual variable in them.

Example:
As you can see from this example, clean_array_gpc() takes 2 paramaters. The first paramater specifies which Superglobal array you wish to clean, and the second is an array of variables and their types.

So, in the example above, we are telling clean_array_gpc() that we wish to clean the $_POST array, and that $_POST contains 3 variables, 'name', 'age', and 'usepm', and that we wish to clean them as TYPE_NOHTML, TYPE_UINT and TYPE_BOOL respectivly.

Once cleaned, the new (clean) variables will be available in the $vbulletin->GPC array. So, to follow on from our previous example, we would use something like:

Cleaning a Single Superglobal Variable

If you have a single variable that you wish to clean, use the clean_gpc() function. This function allows you to specify a single variable in any of the Superglobal arrays, and it's type.

Example:
In this example, the 'age' variable in the $_GET Superglobal array will be cleaned to make sure it is an unsigned integer.

Cleaning a Single Variable

If you wish to clean a single variable that is not in one of the Superglobal arrays, you should use the clean() function.

Example:
From this example you can see that clean() takes 2 paramaters. The first is the variable that you wish to clean and the second is its type. Unlike the last 2 functions, clean() returns the variable directly.

Cleaning an Array of Variables

For times when you wish to clean an array of variables of mixed types, vBulletin provides the clean_array() function. The clean_array() function takes 2 paramaters. The first is the array to be cleaned, and the second is an array of variable names, and their types.

This function works exactly the same as clean_array_gpc(), except instead of specifying which Superglobal array to clean, you specify your own array.

Conclusion

So, to sum up - always run all input from the user through the vBulletin Input Cleaner! As well as being a good coding practice, this will drasticly decrease the chances of someone exploiting your script using an XSS or SQL attack.

Good luck using your new found knowledge of the vBulletin Input Cleaner class, and remember: If you get stuck, just ask! Knowledge sharing is what vBulletin.org is all about!

(Note: If you want to reproduce this article anywhere, I have no objections, but I do request that you give me credit for writing it, and a PM letting me know would be appreciated :))

Another fine article. Good work.

very nice. thanks dude

Thanks Alan, i have been trying to use the cleaners as opposed to $_GET, $_POST, etc... this should help me well.

This is great. One question for a novice hack. How do we know there as something wrong? If I wanted a number and the user input a letter how would I know there was an error?

Do I need to further determine the cleaned variable is a number?

Mark

You'd use
if(!is_numeric($vbulletin->GPC['age']))

For reference, when checking numbers, if vBulletin doesn't find a number, it will return 0 instead. (when using TYPE_INT or TYPE_UINT)

The types and what it will return if it can't find what you are asking for is below:

Thanks,
Alan.

Thanks for the information. So I'm thinking I need to check the variables even before it posts with javascript or with php.

You can always do:

If the value of 0 isn't a possibility.

how would you go about cleaning $_SESSION arrays? I have been cleaning the vars before placing them into $_SESSION but I wonder if there's some way to clean the $_SESSION afterwards.?