Host Header Injection In Wonder CMS - CVE-2017-14523

By- Samrat Das

Hi Readers

Recently in one of my pentest research, I found a CMS " WonderCMS" hosted in github.

Curious to explore its functionalities, I downloaded and set it up in my local system.

After fiddling with the source code, I found that it did not have any kind of security mechanism to filter the host redirections based on the HTTP header. Using this it became possible to perform host header injection attack.

Title of the Vulnerability:  Host Header Injection

Common Vulnerability Scoring System: 7.0

Vulnerability Class: Injection

Technical Details & Description: The application source code is coded in a way which allows arbitrary host header to be defined leading to redirection/ user url manipulation

CVE ID allocated:  CVE-2017-14523

Product & Service Introduction: Wonder CMS 2.3.1

WonderCMS is an open source CMS (Content Management System) built with PHP, jQuery, HTML and CSS (Bootstrap responsive).

WonderCMS doesn't require any configuration and can be simply unzipped and uploaded to your server/hosting provider. All data is saved to a text file (JSON format) which is easily copied, moved, backed up and restored. WonderCMS runs on less than 50 functions and 850 lines of code and 5 files.

Steps to Re-Produce –

1.       Visit the application

2.       Pass any request to the repeater of an interception proxy

3.       Replace the legitimate header with a header url of your choice

4.       I injected www.google.com            

5.       Proceed and you will receive a successful page re-direction to google.com

Exploitation Technique: A attacker can perform cache poisoning/ user redirection and even user to get tricked to visit a malicious site

Severity Level: High

Security Risk:

The presence of such a risk can lead to arbitrary user re-direction

Affected Product Version: 2.3.1

Solution - Fix & Patch: The application code should be configured in such a way, that it should not allow user redirections to arbitrary hosts from the HTTP header and use pre-defined/ sanitized requests.

Step by step for the attack: