IBM WebSphere Application Server Cross-Site Request Forgery

Advisory ID Internal
CORE-2010-1021

1. Advisory Information

Title: IBM WebSphere Application Server Cross-Site Request Forgery
Advisory ID: CORE-2010-1021
Date published: 2011-06-15
Date of last update: 2011-06-15
Vendors contacted: IBM
Release mode: User release

2. Vulnerability Information

Class: Cross-Site Request Forgery (CSRF) [CWE-352]
Impact: Code execution
Remotely Exploitable: Yes
Locally Exploitable: No
CVE Name: CVE-2010-3271

3. Vulnerability Description

WebSphere is IBM's integration software platform. It includes the entire middleware infrastructure --such as servers, services, and tools-- needed to write, run, and monitor 24x7 industrial-strength, on demand Web applications and cross-platform, cross-product solutions. WebSphere Application Server is the base for the infrastructure; everything else runs on top of it [1].

The administrative console of IBM WebSphere Application Server is vulnerable to Cross-Site Request Forgery (CSRF) attacks, which can be exploited by remote attackers to force a logged-in administrator to perform unwanted actions on the IBM WebSphere administrative console, by enticing him to visit a malicious web page.

4. Vulnerable packages

  • IBM WebSphere Application Server 7.0.0.11
  • IBM WebSphere Application Server 7.0.0.13
  • Older versions are probably affected too, but they were not checked.

5. Non-vulnerable packages

Contact the vendor for a fix.

6. Vendor Information, Solutions and Workarounds

Contact the vendor for a fix. The following are workarounds for this issue.

6.1. Server side

According to OWASP [2], CSRF vulnerabilities can be avoided by checking the referrer of the HTTP request and verifying that the request comes from the original site. A potential workaround is thus to set a rule on a Web Application Firewall that checks the referrer of the requests, and verifies that all the requests to the WebSphere administrative console are originated from the same site.

6.2. Client side

An administrator of WebSphere administrative console could mitigate the bug by using Firefox and the NoScript add-on; more precisely by making use of the ABE [3] (Application Boundaries Enforcer) feature of NoScript. With ABE it is possible to define rules such as the following:

Site *.example.com Accept from SELF Deny


This rule applies to *.example.com; it will allow all the requests made from the same site, and block all the requests directed to *.example.com but generated from any other site, avoiding that Firefox sends the request to the server. The syntax of the ABE rules is defined here: http://noscript.net/abe/abe_rules.pdf

7. Credits

This vulnerability was discovered and researched by Francisco Falcon from Core Security Technologies during Bugweek 2010 [4]. Additional research was performed by Alejandro Rodriguez. Publication was coordinated by Carlos Sarraute.

8. Technical Description / Proof of Concept Code

The administrative console (also known as Integrated Solutions Console) of IBM WebSphere Application Server is vulnerable to Cross-Site Request Forgery (CSRF) [2] attacks, which can be exploited by remote attackers to force a logged-in administrator to perform unwanted actions on the IBM WebSphere administrative console, by enticing him to visit a malicious web page.

The administrative console of IBM WebSphere Application Server includes a standard protection mechanism against Cross-Site Request Forgery, which consists of a token that is included as a hidden field on every FORM, named csrfid, that is sent to the web server in each POST request performed by the web browser. When the web server receives a POST request, it checks that the csrfid token included in the parameters of the POST request matches the anti-CSRF token associated with the current session. If they do not match, then IBM WebSphere responds with an "Unauthorized Request" message, thus effectively preventing CSRF.

However, in certain areas of the administrative console, WebSphere forgets to check the value of the csrfid token when processing POST requests, even though the csrfid hidden field is included in every FORM, making the application vulnerable to Cross-Site Request Forgery.

The vulnerable areas of the WebSphere administrative console include the Security > Global Security panel [6], and the Save changes to the master configuration feature. This makes possible for a remote attacker to disable the Administrative Security, Application Security and Java 2 Security options, and then to save the changes to the configuration, by tricking an IBM WebSphere administrator which is currently logged in to the administrative console to visit a malicious web page. Also note that IBM WebSphere 7.0 with Fix Pack 11 did not include a csrfid token for the Save changes to the master configuration feature; Fix Pack 13 introduced it, but anyways it is ignored on the server side when processing a request to save the master configuration.

The following HTML code is a Proof-of-Concept of a specially crafted web page that will leverage the CSRF vulnerability in order to disable the Administrative Security, Application Security and Java 2 Security options, if a logged-in administrator visits it:

<html> <body> <iframe id="iframe1" style="visibility:hidden"></iframe> <iframe id="iframe2" style="visibility:hidden"></iframe> <script> //The first request disables "Administrative security" and "Application security" options document.getElementById("iframe1").src = "https://<ip>:9043/ibm/console/adminSecurityDetail.do?action=Edit&displayActiveUserRegistry=Repositorios+federados&selectUserRegistry=WIM&activeAuthMechanism=LTPA&apply=Aplicar"; //The second request saves the changes in the WebSphere configuration document.getElementById("iframe2").src = "https://<ip>:9043/ibm/console/syncworkspace.do?saveaction=save&directsave=true"; </script> </iframe> </body> </html>

9. Report Timeline

  • 2010-11-24: Core Security contacts IBM, requesting the proper point of contact to report a security vulnerability in IBM WebSphere Application Server.
  • 2010-11-29: Vendor responds providing the point of contact to report the vulnerability, and its PGP key to encrypt communications.
  • 2010-12-14: Core sends an advisory draft, containing the technical details needed to reproduce the vulnerability. Publication of Core's advisory is temporarily set to January 10, 2011.
  • 2010-12-14: IBM acknowledges the receipt of the technical information.
  • 2010-12-21: Core asks the vendor whether it was able to reproduce the vulnerability.
  • 2011-01-05: Vendor responds that it was able to reproduce the issue and confirms there is a vulnerability. Vendor informs Core that it is still working through the total products affected, that it is multiple products, and that this vulnerability is creating real issues on being able to resolve it. Vendor requests Core an extension on the release date while it completes the full evaluation of risk assessment and remedy production. Vendor expects to have that information in the following 2 weeks.
  • 2011-01-06: Core responds that it is willing to postpone the publication of its advisory. However to take that decision more information about the vendor's analysis of the vulnerability and its plans for developing a fix is required. In particular, Core requests a list of all affected products and versions, and also some insight on the difficulties of fixing this issue. In the meantime, the publication of this advisory is rescheduled to February 15th, 2011. (No reply received.)
  • 2011-01-31: Since more than 3 weeks have passed since the last communication, Core requests an update on this issue. In particular Core requests to receive information respect to:
    • the vendor's analysis of the vulnerability,
    • the vendor's plans for developing a fix,
    • a list of affected products and versions.
  • 2011-02-01: Core reminds the vendor that in case of not receiving an answer, it will publish its advisory as "user release" on the scheduled date (February 15th, 2011).
  • 2011-02-01: Vendor replies that it has asked a status update from the WebSphere team to convey to Core, and will provide it briefly.
  • 2011-02-08: Core requests an update on this issue.
  • 2011-02-14: Core reminds the vendor that the advisory is scheduled to be published on February 15th. Core communicates its willingness to publish the advisory as "coordinated release" based on concrete feedback from the vendor.
  • 2011-02-14: Vendor communicates Core that it is working on a statement to provide for Core, and that since the PSIRT is a new mechanism within IBM, it is still defining the way to provide consistent statements. In the meantime, the vendor informs that:
    • The vendor has a potential solution designed and partially implemented to fully secure the console. It is in the process of reviewing the design and the impact to stack products.
    • There are an unknown number of stack products affected. WebSphere Application Server (WAS) stack products that use the ISC (Integrated Solutions Console) based console are affected. The vendor is still gathering the list of products affected, and must determine the impact of implementing the fix.
    • There is a meeting planned to decide on the final solution to be implemented and determine the key delivery dates. These decisions will be taken in mid March or later.
    • The target dates for release reach into Q3 2011.
  • 2011-02-17: Core replies that it has rescheduled publication of its advisory (for the second time) to March 21, 2011, in order to give PSIRT more time to come up with concrete responses to the requested information. Core provides additional information about its own publication process. Without additional information, it is difficult for Core to understand the reason why users of vulnerable WebSphere software should remain without any solution until Q3 2011.
  • 2011-03-17: After 1 month of silence, the vendor informs Core that IBM's point of contact for this issue has changed, and that further communications will be handled by the head of IBM's Secure By Design initiative which includes the IBM PSIRT.
  • 2011-03-17: Vendor requests Core to postpone the publication of its advisory until early October 2011.
  • 2011-03-18: Vendor communicates that since Core hasn't responded to the request (sent the previous day) of deferring the public disclosure of this security vulnerability from 21 March to early October 2011, IBM considers that Core agrees.
  • 2011-03-21: Core answers that October 2011 is well beyond what it considers a reasonable timeframe to patch the type of bug that it has reported (a Cross-Site Request Forgery). Additionally the vendor didn't provide Core a technical analysis of the bug, explaining the difficulty to patch it (and why it would take IBM around 10 months to release fixes). The vendor didn't provide either the requested list of affected products and versions. According to Core's publication policy, the decision of postponing the publication of an advisory cannot be taken without technical arguments that justify that decision. This is why Core cannot agree with IBM's request to postpone publication until October 2011, unless the requested technical information is provided by the vendor. (No reply received.)
  • 2011-04-25: Core communicates the vendor that it has rescheduled the publication of its advisory to June 14th, 2011. That date corresponds to a 6 month timeframe after technical details about this vulnerability were sent to IBM (on December 14th, 2010), and is considered final. (No reply received.)
  • 2011-06-15: The advisory CORE-2010-1021 is published.

10. References

[1] IBM WebSphere Application Server:
http://www-01.ibm.com/software/webservers/appserv/was/

[2] Cross-Site Request Forgery (CSRF)
http://www.owasp.org/index.php/Cross-Site_Request_Forgery_%28CSRF%29

[3] Application Boundaries Enforcer (ABE)
http://noscript.net/abe/

[4] The author participated in Core Security's Bugweek 2010 as member of the team "Ex Tester fuErTes and Exploit Testers".

[5] IBM WebSphere Reference, Global Security settings:
http://publib.boulder.ibm.com/infocenter/wasinfo/fep/topic/com.ibm.websphere.nd.multiplatform.doc/info/ae/ae/usec_secureadminappinfra.html

11. About CoreLabs

CoreLabs, the research center of Core Security Technologies, is charged with anticipating the future needs and requirements for information security technologies. We conduct our research in several important areas of computer security including system vulnerabilities, cyber attack planning and simulation, source code auditing, and cryptography. Our results include problem formalization, identification of vulnerabilities, novel solutions and prototypes for new technologies. CoreLabs regularly publishes security advisories, technical papers, project information and shared software tools for public use at: www.coresecurity.com/core-labs.

12. About Core Security 

Core Security enables organizations to get ahead of threats with security test and measurement solutions that continuously identify and demonstrate real-world exposures to their most critical assets. Our customers can gain real visibility into their security standing, real validation of their security controls, and real metrics to more effectively secure their organizations.

Core Security's software solutions build on over a decade of trusted research and leading-edge threat expertise from the company's Security Consulting Services, CoreLabs and Engineering groups. www.coresecurity.com

13. Disclaimer

The contents of this advisory are copyright (c) 2011 Core Security Technologies and (c) 2011 CoreLabs, and are licensed under a Creative Commons Attribution Non-Commercial Share-Alike 3.0 (United States) License: http://creativecommons.org/licenses/by-nc-sa/3.0/us/