Software Engineering Institute

CWE-79: Improper Neutralization of Input During Web Page Generation ('Cross-site Scripting')

Alfresco Enterprise is vulnerable to a stored cross-site scripting attack when a user uploads an XHTML file. Although uploaded HTML files are validated for javascript code, XHTML files are not. Therefore, a remote authenticated attacker could upload an XHTML file containing arbitrary javascript code. When a victim visits the XHTML file in their browser, the javascript will be executed.

Additionally, the software relies upon an HTML parser to filter untrusted input from various input fields within the application. A specially crafted HTML input allows an authenticated attacker to bypass the HTML parser and persist javascript on various pages of Alfresco. Specifically, the HTML parser considers the <% tag to introduce JSP/ASP code, which the application considers to be unharmful. By escaping crafted input with the <% tag, an attacker can evade the application's anti-XSS mechanisms.

Lastly, the taskId GET parameter to the /share/page/task-edit URL is vulnerable to a reflected cross-site scripting vulnerability.

A remote attacker may be able to execute arbitrary script in the context of the end-user's browser session. With the exception of the vulnerability in /share/page/task-edit, the attacker must be authenticated.

Alfresco has released hotfix 4.1.6.13 to address this issue. Alternatively, users can upgrade to version 4.1.8 or later. In addition, please consider the following workaround:

Restrict access

As a general good security practice, only allow connections from trusted hosts and networks. Note that restricting access does not prevent XSS attacks since the attack comes as an HTTP request from a legitimate user's host. Restricting access would prevent an attacker from accessing the web interface using stolen credentials from a blocked network location.