Infection Monkey

Data center Security Tool

http://www.guardicore.com/the-infected-chaos-monkey/

The Infection Monkey is an open source security tool for testing a data center's resiliency to perimeter breaches and internal server infection. The Monkey uses various methods to self propagate across a data center and reports success to a centralized C&C server.

Features include:

• Multiple propagation techniques:
  • Predefined passwords
  • Common exploits
• Multiple propagation protocols:
  • SSH
  • SMB
  • RDP
• A C&C server with a dedicated UI to visualize the Monkey's progress inside the data center

Getting Started

The Infection Monkey is comprised of two parts: the Monkey and the C&C server. The monkey is the tool which infects other machines and propagates to them, while the C&C server collects all Monkey reports and displays them to the user.

Requirements

The C&C Server has been tested on Ubuntu 14.04. The Monkey itself has been tested on Windows XP, 7, 8.1 and 10. The Linux build has been tested on Ubuntu server 14.04 and 15.10.

Installation

For off-the-shelf use, download our pre-compiled binaries from our website. To set up the C&C server follow the instructions on Monkey Island readme. If you wish to compile the binaries yourself, follow the instructions under Building the Monkey from Source.

Initial configuration.

Whether you're downloading or building the Monkey from source, the Infection Monkey is comprised of 4 executable files for different platforms plus a default configuration file.

Monkey configuration is stored in two places:

1. By default, the Monkey uses a local configuration file (usually, config.bin). This configuration file must include the address of the Monkey's C&C server.
2. After successfully connecting to the C&C server, the monkey downloads a new configuration from the server and discards the local configuration. It is possible to change the default configuration from the C&C server's UI.

In both cases the command server hostname should be modified to point to your local instance of the Monkey Island (note that this doesn't require connectivity right off the bat). In addition, to improve the Monkey's chances of spreading, you can pre-seed it with credentials and usernames commonly used.

Both configuration options use a JSON format for specifying options; see "Options" below for details.

Running the C&C Server

To run the C&C Server, install our infected Monkey debian package on a specific server. The initial infected machine doesn't require a direct link to this server.

Unleashing the Monkey

Once configured, run the monkey using ./monkey-linux-64 m0nk3y -c config.bin -s 41.50.73.31:5000 (Windows is identical). This can be done at multiple points in the network simultaneously.

Command line options include:

• -c, --config: set configuration file. JSON file with configuration values, will override compiled configuration.
• -p, --parent: set monkey’s parent uuid, allows better recognition of exploited monkeys in c&c
• -t, --tunnel: ip:port, set default tunnel for Monkey when connecting to c&c.
• -d, --depth : sets the Monkey's current operation depth.

How to activate the Monley

1. Wakeup connection to c&c, sends basic info of the current machine and the configuration the monkey uses to the c&c.
2. First try direct connection to c&c.
3. If direct connection fails, try connection through a tunnel, a tunnel is found according to specified parameter (the default tunnel) or by sending a multicast query and waiting for another monkey to answer.
4. If no connection can be made to c&c, continue without it.
5. If a firewall app is running on the machine (supports Windows Firewall for Win XP and Windows Advanced Firewall for Win 7+), try to add a rule to allow all our traffic.
6. Startup of tunnel for other Monkeys (if connection to c&c works).
7. Firewall is checked to allow listening sockets (if we failed to add a rule to Windows firewall for example, the tunnel will not be created)
8. Will answer multicast requests from other Monkeys in search of a tunnel.
9. Running exploitation sessions, will run x sessions according to configuration:
10. Connect to c&c and get the latest configuration
11. Scan ip ranges according to configuration.
12. Try fingerprinting each host that answers, using the classes defined in the configuration (SMBFinger, SSHFinger, etc)
13. Try exploitation on each host found, for each exploit class in configuration: 1. check exploit class supports target host (can be disabled by configuration) 2. each exploitation class will use the data acquired in fingerprinting, or during the exploit, to find the suitable Monkey executable for the host from the c&c.
    1. If c&c connection fails, and the source monkey’s executable is suitable, we use it.
    2. If a suitable executable isn’t found, exploitation will fail.
    3. Executables are cached in memory.
14. will skip hosts that are already exploited in next run
15. will skip hosts that failed during exploitation in next run (can be disabled by configuration)
16. Close tunnel before exiting Wait for monkeys using the tunnel to unregister for it Cleanup Remove firewall rules if added

Configuration Options

Key	Type	Description	Possible Values
singleton_mutex_name	string	string of the mutex name for single instance	example: {2384ec59-0df8-4ab9-918c-843740924a28}
alive	bool	sets whether or not the monkey is alive. if false will stop scanning and exploiting
self_delete_in_cleanup	bool	sets whether or not to self delete the monkey executable when stopped
use_file_logging	bool	sets whether or not to use a log file
monkey_log_path_[windows/linux]	string	file path for monkey logger.
timeout_between_iterations	int	how long to wait between scan iterations
max_iterations	int	how many scan iterations to perform on each run
victims_max_find	int	how many victims to look for in a single scan iteration
victims_max_exploit	int	how many victims to exploit before stopping
command_servers	array	addresses of c&c servers to try to connect	example: ["russian-mail-brides.com:5000"]
internet_services	array	addresses of internet servers to ping and check if the monkey has internet acccess
retry_failed_explotation	bool	sets whether or not to retry failed hosts on next scan
range_class	class name	sets which ip ranges class is used to construct the list of ips to scan	FixedRange - scan list is a static ips list, RelativeRange - scan list will be constructed according to ip address of the machine and size of the scan, ClassCRange - will scan the entire class c the machine is in.
scanner_class	class name	sets which scan class to use when scanning for hosts to exploit	TCPScanner - searches for hosts according to open tcp ports, PingScanner - searches for hosts according to ping scan
finger_classes	tuple of class names	sets which fingerprinting classes to use	in the list: SMBFinger - get host os info by checking smb info, SSHFinger - get host os info by checking ssh banner, PingScanner - get host os type by checking ping ttl. For example: (SMBFinger, SSHFinger, PingScanner)
exploiter_classes	tuple of class names		SmbExploiter - exploit using smb connection, WmiExploiter - exploit using wmi connection, RdpExploiter - exploit using rdp connection, Ms08_067_Exploiter - exploit using ms08_067 smb exploit, SSHExploiter - exploit using ssh connection
range_fixed	tuple of strings	list of ips to scan
RelativeRange range_size	int	number of hosts to scan in relative range
tcp_target_ports	list of int	which ports to scan using TCPScanner
tcp_scan_timeout	int	timeout for tcp connection in tcp scan (in milliseconds)
tcp_scan_interval	int	time to wait between ports in the tcp scan (in milliseconds)
tcp_scan_get_banner	bool	sets whether or not to read a banner from the tcp ports when scanning
ping_scan_timeout	int	timeout for the ping command (in milliseconds) utilised by PingScanner
psexec_user	string	user to use for connection, utilised by SmbExploiter/WmiExploiter/RdpExploiter
psexec_passwords	list of strings	list of passwords to use when trying to exploit
skip_exploit_if_file_exist	bool	sets whether or not to abort exploit if the monkey already exists in target, used by SmbExploiter
rdp_use_vbs_download	bool	sets whether to use vbs payload for rdp exploitation in RdpExploiter. If false, bits payload is used (will fail if bitsadmin.exe doesn’t exist)
ms08_067_exploit_attempt	int	number of times to try and exploit using ms08_067 exploit
ms08_067_remote_user_add	string	user to add to target when using ms08_067 exploit
ms08_067_remote_user_pass	string	password of the user the exploit will add
ssh_user	string	user to use for ssh connection, used by SSHExploiter
ssh_passwords	list of strings	list of passwords to use when trying to exploit using SSHExploiter

Building the Monkey from source

If you want to build the monkey from source instead of using our provided packages, follow the instructions at the readme files under chaos_monkey and monkey_island.

License

Copyright (c) 2016 Guardicore Ltd

See the LICENSE file for license rights and limitations (GPLv3).

Dependent packages

Dependency	License
libffi-dev	https://github.com/atgreen/libffi/blob/master/LICENSE
PyCrypto	Public domain
upx	Custom license, http://upx.sourceforge.net/upx-license.html
bson	BSD
enum34	BSD
pyasn1	BSD
psutil	BSD
flask	BSD
flask-Pymongo	BSD
Flask-Restful	BSD
python-dateutil	Simplified BSD
zope	ZPL 2.1
Bootstrap	MIT
Bootstrap Switch	Apache 2.0
Bootstrap Dialog	MIT
JSON Editor	MIT
Datatables	MIT
jQuery	MIT
cffi	MIT
twisted	MIT
typeahead.js	MIT
Font Awesome	MIT
vis.js	MIT/Apache 2.0
impacket	Apache Modified
Start Bootstrap (UI Theme)	Apache 2.0
requests	Apache 2.0
odict	Python Software Foundation License
paramiko	LGPL
rdpy	GPL-3
winbind	GPL-3
pyinstaller	GPL
Celery	BSD