Python low-interaction honeyclient

Last update: Dec 19, 2022

Overview

Thug

The number of client-side attacks has grown significantly in the past few years shifting focus on poorly protected vulnerable clients. Just as the most known honeypot technologies enable research into server-side attacks, honeyclients allow the study of client-side attacks.

A complement to honeypots, a honeyclient is a tool designed to mimic the behavior of a user-driven network client application, such as a web browser, and be exploited by an attacker's content.

Thug is a Python low-interaction honeyclient aimed at mimicing the behavior of a web browser in order to detect and emulate malicious contents.

Documentation

Documentation about Thug installation and usage can be found at http://thug-honeyclient.readthedocs.io/.

Contributions

Thug is open source and we welcome contributions in all forms!

Thug is free to use for any purpose (even commercial ones). If you use and appreciate Thug, please consider supporting the project with a donation using Paypal.

Testing

To run the full test suite using tox, run the command:

tox

Since tox builds and installs dependencies from scratch, using pytest for faster testing is recommended:

pytest --cov thug

License information

License: GNU General Public License, version 2

Comments

Add support for WScriptNetwork object.

Adds support for the WScriptNetwork object which was used in some recent samples.

Ideally the attributes shouldn't be static but retrieved using WScriptShell.ExpandEnvironmentStrings(). Any comments on how to do that?

Also did some improvements on the Scripting.FileSystemObject module.
enhancement

opened by tweemeterjop 17

Thug not properly handling redirects logging

Hi! Let's take a look at this situation: At hxxp://localhost:1337/ek_tests/dummy_redirect_chain/exploit.html some code was added by nasty hacker.

<html>
<head>
<title>Dummy sploit</title>
</head>
<body>
<!-- malicious code start -->
<script src="javascript.php?option=0"></script>
<!-- malicious code end -->
</body>
</html>

That code redirects (indirectly) to (imaginated) exploit kit's gate (that uses alert(1) from js as exploit). It creates some new scripts on the way to gate. At the backend side, it looks like this:

<?php
/**
* javascript.php
*/
$i = $_GET['option'];
$to_write = "";
switch ($i) {
    case 0:
        $to_write =  "var script = document.createElement('script');
script.src = 'javascript.php?option=1';
document.body.appendChild(script);";
        break;
    case 1:
        header('Location: javascript.php?option=2');
        break;
    case 2:
        $to_write = "document.write(\"<iframe src='iframe.php?option=1'></iframe>\");";
        break;
   case 3:
        $to_write =  "var script = document.createElement('script');
script.src = 'javascript.php?option=4';";
    case 4:
       $to_write = "alert('1');";
}
echo $to_write;
?>

Also uses iframes

<?php
/**
* iframe.php
*/
$i = $_GET['option'];
$to_write = "";
switch ($i) {
    case 0:
        $to_write =  "var script = document.createElement('script');
script.src = 'javascript.php?option=1';";
        break;
    case 1:
        $to_write = "<script src='javascript.php?option=3'></script>";
        break;
    case 2:
        $to_write = "document.write(\"<iframe src='iframe.php?option=1'></iframe>\");";
        break;
   case 3:
        $to_write =  "var script = document.createElement('script');
script.src = 'javascript.php?option=4';";
    case 4:
       $to_write = "document.alert(1);";
}
echo '<html><head></head><body>';
echo $to_write;
echo '</body></html>';
?>

When user (thug) enters that site, it produces some redirects with alert(1) at the end. It looks like this: http://svgur.com/i/1XW.svg - graph made by thug itself.

The problem is: Even if one script leads to another, we are not informed about that. We cannot build proper redirect chains for that redirections. We get such output:

 | -> 1    / (2->3 redirected by HTTP Location header)
0| -> 2 -> 3
 | -> 4

Even if that really should be:

0 -> 1 -> 2 -> 3 -> 4

I know that this is caused by, for example, that newly created scripts are executed within base (0's) DOM. I believe we could somehow get information about real source of some action from V8. It is critical aspect for analysis of web threats, because at the moment we don't really know what part of website is malicious, and what is not. Also, thanks for a great tool :)

enhancement

opened by internaldefect 16

new version - missing graph.svg

Using thug via docker installation, very good tool. Recently installed the new version but it seems the graph.svg file is not being created within the analysis folder.

Is this feature removed completely or can it be enabled?

Thanks
configuration

opened by arunkhan1984 13

First time set-up problems.

Hi, firstly I am on Ubuntu 16.04 and have gone through the Thug installation guide multiple time and am currently stuck as where to go from here. I installed all packages except HoneyAgent & VirusTotal (both optional). My problem is that when I run thug --list-ua I get this:


[2017-03-28 10:11:03] [HTML Classifier] Skipping not existing default classification rule file
[2017-03-28 10:11:03] [HTML Classifier] Skipping not existing default filter file
[2017-03-28 10:11:03] [JS Classifier] Skipping not existing default classification rule file
[2017-03-28 10:11:03] [JS Classifier] Skipping not existing default filter file
[2017-03-28 10:11:03] [VBS Classifier] Skipping not existing default classification rule file
[2017-03-28 10:11:03] [VBS Classifier] Skipping not existing default filter file
[2017-03-28 10:11:03] [URL Classifier] Skipping not existing default classification rule file
[2017-03-28 10:11:03] [URL Classifier] Skipping not existing default filter file
[2017-03-28 10:11:03] [Sample Classifier] Skipping not existing default classification rule file
[2017-03-28 10:11:03] [Sample Classifier] Skipping not existing default filter file
[2017-03-28 10:11:03] [Text Classifier] Skipping not existing default classification rule file
[2017-03-28 10:11:03] [Text Classifier] Skipping not existing default filter file

Synopsis:
    Thug: Pure Python honeyclient implementation

    Available User-Agents:

So lots of skipping, I don't know if that is a problem but no User-Agents? Also when I try to run Thug I get the following:

[2017-03-28 10:15:41] [HTML Classifier] Skipping not existing default classification rule file
[2017-03-28 10:15:41] [HTML Classifier] Skipping not existing default filter file
[2017-03-28 10:15:41] [JS Classifier] Skipping not existing default classification rule file
[2017-03-28 10:15:41] [JS Classifier] Skipping not existing default filter file
[2017-03-28 10:15:41] [VBS Classifier] Skipping not existing default classification rule file
[2017-03-28 10:15:41] [VBS Classifier] Skipping not existing default filter file
[2017-03-28 10:15:41] [URL Classifier] Skipping not existing default classification rule file
[2017-03-28 10:15:41] [URL Classifier] Skipping not existing default filter file
[2017-03-28 10:15:41] [Sample Classifier] Skipping not existing default classification rule file
[2017-03-28 10:15:41] [Sample Classifier] Skipping not existing default filter file
[2017-03-28 10:15:41] [Text Classifier] Skipping not existing default classification rule file
[2017-03-28 10:15:41] [Text Classifier] Skipping not existing default filter file
[2017-03-28 10:15:41] [WARNING] VirusTotal disabled (no configuration file found)
[2017-03-28 10:15:41] [CRITICAL] Logging subsystem not initialized (configuration file not found)
Traceback (most recent call last):
  File "/usr/local/bin/thug", line 11, in <module>
    sys.exit(main())
  File "/usr/local/lib/python2.7/dist-packages/thug/thug.py", line 334, in main
    Thug(sys.argv[1:])()
  File "/usr/local/lib/python2.7/dist-packages/thug/ThugAPI/ThugAPI.py", line 86, in __call__
    self.analyze()
  File "/usr/local/lib/python2.7/dist-packages/thug/thug.py", line 324, in analyze
    ThugPlugins(PRE_ANALYSIS_PLUGINS, self)()
  File "/usr/local/lib/python2.7/dist-packages/thug/Plugins/ThugPlugins.py", line 44, in __init__
    self.get_plugins()
  File "/usr/local/lib/python2.7/dist-packages/thug/Plugins/ThugPlugins.py", line 68, in get_plugins
    for p in os.listdir(PLUGINS_PATH):
OSError: [Errno 2] No such file or directory: '/etc/thug/plugins'

I have followed the installation guide step by step and it is now at the Usage section? Am I meant to install the Docker too or is there something else I have missed?

Thank you for your help in advance!

opened by Waffles-2 11

ReferenceError

There is malicious code and normal code in javascript. If a reference error occurs in the normal code, the malicious code will not be analyzed in the future. Can I pass(skip) the reference error?... and then I want to continue analyzing the malicious code

[2018-11-12 21:47:41] Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/thug-0.9.37-py2.7.egg/thug/Debugger/Shellcode.py", line 162, in run result = self.ctxt.eval (self.script) ReferenceError: ReferenceError: functions is not defined (@ 1: 0) -> functions.validateAndSubmitForm ();
not an issue

opened by parknom 10
Data in mongodb

Hi, I wonder how possible it would be to push most of the data in analysis.xml/avlog.json and graph.svg into mongodb so it can be used by eg. a web interface.

I really want to use thug, but I find the lack of a web interface for viewing the reports quite the problem.

Ideally I guess a JSON version of the MAEC schema would be the best, but that kinda goes against the MAEC specifications, for now.. :)

An example of a web interface I could make with having all the data in mongodb could be something like http://two.mrfjo.org/02906d5b-a7cd-47f8-9885-6e80eecf967a/
enhancement

opened by espenfjo 10
Fix infinite recursion problem in Window.__getattr__
Window object has infinite recursion when the following two conditions:

Check whether the V8 JSContext has been initialized through getattr(). It would be a lookup loop which always calls getattr() since the code uses self.context before initializing V8 JSContext.

Fix by: Raise attribute error if getattribute() can't find "_context" variable.

Get the nonexistent variable or function by "self.context.eval(name)". When V8 JSContext can't find the variable, it will call getattr() again. It also becomes a lookup loop.

Fix by: Avoid to call getattr() from V8Context internal.

> python -m cProfile thug.py -l ../samples/Events/testEvent10.html

Before fix: ncalls tottime percall cumtime percall filename:lineno(function) 588/5 0.193 0.000 0.759 0.152 Window.py:138(getattr) 6825 0.003 0.000 0.003 0.000 Window.py:196(window) 616/344 0.378 0.001 0.599 0.002 Window.py:850(context)

After fix: ncalls tottime percall cumtime percall filename:lineno(function) 55/37 0.014 0.000 0.014 0.000 Window.py:139(getattr) 35 0.000 0.000 0.000 0.000 Window.py:206(window) 82/4 0.004 0.000 0.020 0.005 Window.py:860(context)
opened by Joseph-CHC 10

javascript that fools thug

Saw this on a landing page, thug did not follow it:

<script type="text/javascript">
    setTimeout("testTime()", 0);
    function testTime() {
        location = "/";
    }
 </script>

defect

opened by evilscheme 10

ImportError: libemu.so.2: cannot open shared object file: No such file or directory

I have successfully installed V8 and PyV8. i have tested it and the result is OK.

but while i was executing python thug.py -h,, the error was appeared. the errors look like:

[email protected]:/home/kafin/thug/src# sudo python thug.py -h Traceback (most recent call last): File "thug.py", line 31, in from DOM import Window, DFT, MIMEHandler, SchemeHandler File "/home/kafin/thug/src/DOM/Window.py", line 32, in from .Location import Location File "/home/kafin/thug/src/DOM/Location.py", line 22, in import DFT File "/home/kafin/thug/src/DOM/DFT.py", line 20, in import pylibemu ImportError: libemu.so.2: cannot open shared object file: No such file or directory

anyone got also this error, or how do i solve this?

thanks

opened by kafin 10
Problem with shellcode encoding - MongoDB logging
Hello, I think that the shellcode is not properly handled and stored in MongoDB. I get error messages like this one when thug tries to decode the hex encoded shellcode:

... File "/usr/local/lib/python2.7/dist-packages/thug/Logging/ThugLogging.py", line 125, in add_shellcode_snippet m(snippet, language, relationship, tag.hex, method) File "/usr/local/lib/python2.7/dist-packages/thug/Logging/modules/MongoDB.py", line 416, in add_code_snippet 'snippet' : self.fix(snippet), File "/usr/local/lib/python2.7/dist-packages/thug/Logging/modules/MongoDB.py", line 408, in fix return thug_unicode(data).replace("\n", "").strip() UnicodeDecodeError: 'ascii' codec can't decode byte 0xeb in position 4: ordinal not in range(128)

The problem is that the shellcode contains non printable characters and thus can not be stored in decoded form. Also even when it is possible to decode the shellcode from unicode encoding the output is not as useful as the original hex encoded form. This exception also breaks the analysis and thus it is not properly finished. I think that the shellcode should not be decoded and should be stored in its original form. What do you think about that? Thank you.

Stanislav
defect
opened by standa4 9
Timeout option not working properly

Hello Is it possible that the timeout option (-T) is not working as intended? I was scanning some targets that take a long time or never finish at all (longer than the 600s default timeout), where I noticed that the runs actually don't finish after the default timeout. I still see the timeout message in the log: [2019-03-19 09:09:30] The analysis took more than 60 second(s). Aborting! however the run is not stopped as result, rather it continues as if nothing happened. The same behaviour was occurring after choosing different targets and a shorter timeout.

I am running thug inside a docker container and am starting the scans with the following command (short timeout for testing purposes): docker run --rm -it buffer/thug thug -T 10 -d <target url>

I am misinterpreting what the timeout option is for? I am assuming it should abort the scan as soon as the timeout is reached, but maybe that is not the case.
defect docker

opened by PinkSheep 8

Releases(v4.7)

v4.7(Jan 4, 2023)

Source code(tar.gz)
Source code(zip)
v4.6.1(Jan 3, 2023)

Source code(tar.gz)
Source code(zip)
v4.6(Dec 28, 2022)

Source code(tar.gz)
Source code(zip)
v4.5(Dec 6, 2022)

Source code(tar.gz)
Source code(zip)
v4.4(Nov 17, 2022)

Source code(tar.gz)
Source code(zip)
v4.3(Nov 10, 2022)

Source code(tar.gz)
Source code(zip)
v4.2(Oct 24, 2022)

Source code(tar.gz)
Source code(zip)
v4.1(Aug 23, 2022)

Source code(tar.gz)
Source code(zip)
v4.0(Aug 4, 2022)

Source code(tar.gz)
Source code(zip)
v3.14(Jul 6, 2022)

Source code(tar.gz)
Source code(zip)
v3.13(Jun 8, 2022)

Source code(tar.gz)
Source code(zip)
v3.12.6(May 31, 2022)

Source code(tar.gz)
Source code(zip)
v3.12.5(May 23, 2022)

Source code(tar.gz)
Source code(zip)
v3.12.4(May 17, 2022)

Source code(tar.gz)
Source code(zip)
v3.12.3(May 9, 2022)

Source code(tar.gz)
Source code(zip)
v3.12.2(May 3, 2022)

Source code(tar.gz)
Source code(zip)
v3.12.1(Apr 29, 2022)

Source code(tar.gz)
Source code(zip)
v3.12(Apr 29, 2022)

Source code(tar.gz)
Source code(zip)
v3.11(Apr 25, 2022)

Source code(tar.gz)
Source code(zip)
v3.10(Apr 6, 2022)

Source code(tar.gz)
Source code(zip)
v3.9(Feb 25, 2022)

Source code(tar.gz)
Source code(zip)
v3.8(Feb 15, 2022)

Source code(tar.gz)
Source code(zip)
v3.7(Jan 24, 2022)

Source code(tar.gz)
Source code(zip)
v3.6.2(Jan 7, 2022)

Source code(tar.gz)
Source code(zip)
v3.6.1(Dec 14, 2021)

Source code(tar.gz)
Source code(zip)
v3.6(Dec 7, 2021)

Source code(tar.gz)
Source code(zip)
v3.5(Oct 13, 2021)

Source code(tar.gz)
Source code(zip)
v3.4(Sep 28, 2021)

Source code(tar.gz)
Source code(zip)
v3.3(Sep 15, 2021)

Source code(tar.gz)
Source code(zip)
v3.2(Sep 9, 2021)

Source code(tar.gz)
Source code(zip)

Owner

Angelo Dell'Aera

Linux kernel hacker, deeply interested in every facet of computer and information security, intensely involved in exploring and analyzing new emerging threats.

GitHub Repository

Python low-interaction honeyclient

Related tags

Overview

Thug

Documentation

Contributions

Testing

License information

Comments

> python -m cProfile thug.py -l ../samples/Events/testEvent10.html

Releases(v4.7)

v4.7(Jan 4, 2023)

v4.6.1(Jan 3, 2023)

v4.6(Dec 28, 2022)

v4.5(Dec 6, 2022)

v4.4(Nov 17, 2022)

v4.3(Nov 10, 2022)

v4.2(Oct 24, 2022)

v4.1(Aug 23, 2022)

v4.0(Aug 4, 2022)

v3.14(Jul 6, 2022)

v3.13(Jun 8, 2022)

v3.12.6(May 31, 2022)

v3.12.5(May 23, 2022)

v3.12.4(May 17, 2022)

v3.12.3(May 9, 2022)

v3.12.2(May 3, 2022)

v3.12.1(Apr 29, 2022)

v3.12(Apr 29, 2022)

v3.11(Apr 25, 2022)

v3.10(Apr 6, 2022)

v3.9(Feb 25, 2022)

v3.8(Feb 15, 2022)

v3.7(Jan 24, 2022)

v3.6.2(Jan 7, 2022)

v3.6.1(Dec 14, 2021)

v3.6(Dec 7, 2021)

v3.5(Oct 13, 2021)

v3.4(Sep 28, 2021)

v3.3(Sep 15, 2021)

v3.2(Sep 9, 2021)

Owner

Angelo Dell'Aera

A Feature Rich Modular Malware Configuration Extraction Utility for MalDuck

Simplify getting and using cookies from the browser to use in Python.

Monty Hall Problem simulation written in Python.

CVE-2021-21972

A Safer PoC for CVE-2022-22965 (Spring4Shell)

A TCP Backdoor made in python

Moodle community-based vulnerability scanner

Show apps recorded storage files by jailbreak

A brute force tool for password-protected zip file

Community Repository for Unofficial Saltbox Add-ons

Some Attacks of Exchange SSRF ProxyLogon&ProxyShell

orfipy is a tool written in python/cython to extract ORFs in an extremely and fast and flexible manner

adb - A tool that allows you to search for vulnerable android devices across the world and exploit them.

A Proof-Of-Concept for the recently found CVE-2021-44228 vulnerability

🔐 A simple command-line password manager.

Apache Flink 目录遍历漏洞批量检测 (CVE-2020-17519)

Unsafe Twig processing of static pages leading to RCE in Grav CMS 1.7.10

WebLogic T3/IIOP RCE ExternalizableHelper.class of coherence.jar

Exploiting CVE-2021-42278 and CVE-2021-42287 to impersonate DA from standard domain user

Ensure secure infrastructure and consistency with the firewall rules