Netgear Voice Gateway - Multiple Vulnerabilities








# Exploit Title: [Netgear Voice Gateway Multiple Vulnerabilities]
# Date: May 01, 2015 [No response from Vendor]
# Discovered by: Karn Ganeshen
# Vendor Homepage: []
# Version: [Firmware Version: V2.3.0.23_2.3.23]

*Netgear Voice Gateway Multiple Vulnerabilities *

*Device Info *
Device Type: Netgear Voice Gateway EVG2000
Account Name: EVG2000
Firmware Version: V2.3.0.23_2.3.23

*1. Web application vulnerabilities OS Command Injection *

Netgear Voice Gateway EVG2000 is managed through a web management portal.
The application provides a Diagnostics feature that has four (4) options:

a.Ping an IP address
b.Perform a DNS Lookup
c.Display the Routing Table
d.Reboot the Router

Option 1 Ping an IP address was confirmed to be vulnerable to OS Command

The ping_IPAddr parameter does not sufficiently validate input. It is
possible to use the semi-colon character (;) to inject arbitrary OS
commands and retrieve the output in the application's responses.


*HTTP POST Request*

POST /ping.cgi HTTP/1.1
User-Agent: blah
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language: en-US
Authorization: Basic <b64_value_here>
Content-Length: 69


*HTTP Response*
root:<redacted_hash>:0:0:Linux User,,,:/root/:/bin/sh

*2. Web application vulnerabilities Stored Cross-Site Scripting (XSS) *

In the Services menu, the Service Table lists any existing Service-Port
mappings. A new service can be added with a payload value of
*<script>alert(xss)</script>* in the ServiceType parameter.

The application does not check any malicious input and accepted this new
entry. The JavaScript input was then returned unmodified in a subsequent
request for the Services Table Entries.

The web application lacks strict input validation and hence is vulnerable
to Stored Cross-Site Scripting attack.

*3. Application does not secure configured passwords (HTTP) *

Any & all configured sensitive information such as passwords & keys are not
secured properly. These are masked and only ***** is shown in the
corresponding fields.

This client-side restriction can easily be bypassed though. It is possible
to capture masked values via ‘Inspect Element’ and / or via an intercepting

The application should mask/censure (*****) the passwords, keys and any
other crucial pieces of configuration and must not pass the values in