Directory traversal attack


A directory traversal consists in exploiting insufficient security validation / sanitization of user-supplied input file names, such that characters representing "traverse to parent directory" are passed through to the file APIs.
The goal of this attack is to use an affected application to gain unauthorized access to the file system. This attack exploits a lack of security as opposed to exploiting a bug in the code.
Directory traversal is also known as the ../ attack, directory climbing, and backtracking. Some forms of this attack are also canonicalization attacks.

Example

A typical example of vulnerable application in PHP code is:

$template = 'red.php';
if
include ;
?>

An attack against this system could be to send the following HTTP request:

GET /vulnerable.php HTTP/1.0
Cookie: TEMPLATE=../../../../../../../../../etc/passwd

Generating a server response such as:

HTTP/1.0 200 OK
Content-Type: text/html
Server: Apache
root:fi3sED95ibqR6:0:1:System Operator:/:/bin/ksh
daemon:*:1:1::/tmp:
phpguru:f8fk3j1OIf31.:182:100:Developer:/home/users/phpguru/:/bin/csh

The repeated ../ characters after /home/users/phpguru/templates/ has caused
to traverse to the root directory, and then include the Unix password file /etc/passwd.
Unix /etc/passwd is a common file used to demonstrate directory traversal, as it is often used by crackers to try cracking the passwords.
However, in more recent Unix systems, the passwd file does not contain the hashed passwords. They are, instead, located in the shadow file which cannot be read by unprivileged users on the machine. It is however, still useful for account enumeration on the machine, as it still displays the user accounts on the system.

Variations of directory traversal

Listed below are some known directory traversal attack strings:

Directory traversal on Unix

Common Unix-like directory traversal uses the ../ characters.
Sudo, a privilege management program ubiquitous in Unix is vulnerable to this attack when users use the glob wildcard.

Directory traversal on Microsoft Windows

and DOS directory traversal uses the ..\ or ../ character sequences.
Each partition has a separate root directory and there is no common root directory above that. This means that for most directory vulnerabilities on Windows, the attack is limited to a single partition.
This kind of attack has been the cause of numerous Microsoft vulnerabilities.

URI encoded directory traversal

problem.
Some web applications scan query string for dangerous characters such as:
to prevent directory traversal. However, the query string is usually URI decoded before use. Therefore, these applications are vulnerable to percent encoded directory traversal such as:
problem.
UTF-8 was noted as a source of vulnerabilities and attack vectors by Bruce Schneier and Jeffrey Streifling.
When Microsoft added Unicode support to their Web server, a new way of encoding ../ was introduced into their code, causing their attempts at directory traversal prevention to be circumvented.
Multiple percent encodings, such as
translated into / or \ characters.
Percent encodings were decoded into the corresponding 8-bit characters by Microsoft webserver. This has historically been correct behavior as Windows and DOS traditionally used canonical 8-bit characters sets based upon ASCII.
However, the original UTF-8 was not canonical, and several strings were now string encodings translatable into the same string. Microsoft performed the anti-traversal checks without UTF-8 canonicalization, and therefore not noticing that C0AF and 2F were the same character when doing string comparisons. Malformed percent encodings, such as %c0%9v was also utilized.

Zip/archive traversal attacks

The use of archive formats like zip allows for directory traversal attacks: files in the archive can be written such that they overwrite files on the filesystem by backtracking. Code that uncompresses archive files can be written to check that the paths of the files in the archive do not engage in path traversal.

Possible methods to prevent directory traversal

A possible algorithm for preventing directory traversal would be to:
  1. Process URI requests that do not result in a file request, e.g., executing a hook into user code, before continuing below.
  2. When a URI request for a file/directory is to be made, build a full path to the file/directory if it exists, and normalize all characters.
  3. It is assumed that a 'Document Root' fully qualified, normalized, path is known, and this string has a length N. Assume that no files outside this directory can be served.
  4. Ensure that the first N characters of the fully qualified path to the requested file is exactly the same as the 'Document Root'.
  5. If so, allow the file to be returned.
  6. If not, return an error, since the request is clearly out of bounds from what the web-server should be allowed to serve.
  7. Using a hard-coded predefined file extension to suffix the path does not limit the scope of the attack to files of that file extension.
include;
The user can use the NULL character in order to bypass everything after the $_GET.

Resources

*
*