Skip to content
forked from Den1al/JSShell

An interactive multi-user web JS shell

License

phoenixml/JSShell

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

95 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

JSShell

An interactive multi-user web based shell written in Python with Flask (for server side) and of course Javascript and HTML (client side). It was initially created in order to debug remote esoteric browsers during tests and research. I am well aware of other purposes this tool might serve, use it at your own responsibility and risk.

Version: 0.4

Author

Daniel Abeles.

Shell Video

asciicast

Installation

It is recommended to use a virtual environment (I used python 3.6, but eariler work just fine):

Pyenv

pyenv virtualenv -p python3.6 venv
pyenv activate venv

virtualenv

virtualenv -p python3.6 venv
source venv/bin/activate

Then for both

pip install -r requirements.txt

Features

  • Multi client support
  • Cyclic DOM objects support
  • Pre flight scripts
  • Command Queue
  • Command Context
  • Injectable via <script> tags
  • Dumping output to file
  • Shell pagination

Running

Create the configuration file

cd app
cp config.py.template config.py

Now you can change the settings you need

Create the database

python db_handler.py create

Start the server (at the background):

python run.py

Navigate with a browser to the server address

If you running locally, then navigate to http://localhost:5000 (port/host can be changed)

Open the interactive shell

python shell.py

Optional : Pre flight scripts

Those are scripts that will execute on every registration of a new client. By default I included pre-flight scripts which grab the following:

  1. Window object
  2. Document object
  3. The browsers screen data
  4. The browser plugins

They are mainly useful when you are automating this process and you know apriori what you want to collect.

Usage

The shell interface contains various commands (can be revealed using the help command).

  ╦╔═╗╔═╗┬ ┬┌─┐┬  ┬
  ║╚═╗╚═╗├─┤├┤ │  │
 ╚╝╚═╝╚═╝┴ ┴└─┘┴─┘┴─┘
  By @Daniel_Abeles

>> help
+-------------+----------------------------------------------------------+
| command     | description                                              |
+-------------+----------------------------------------------------------+
| list        | Lists all the clients registered                         |
| help        | self.help()                                              |
| select <id> | Selected a specific client from the list                 |
| info <id>   | Prints information on a specific client                  |
| <command>   | Executes a command to the current selected client        |
| back        | Detaches from the current client                         |
| exit        | Exists this interactive shell                            |
| coms        | Displays the commands and output for the current client  |
| com <id>    | Displays a specific command and output                   |
| more <id>   | Displays a specific command and output (with pagination) |
| comk        | Kills a command ("*" for all)                            |
| clik        | Kills a client ("*" for all)                             |
| dump <id>   | Dumps the command output to disk - "dump.txt"            |
+-------------+----------------------------------------------------------+                              

Utilizing the command queue, you can fire multiple commands and the client will execute them one by one. All the commands are executed using a single context, so you issue multiple related commands.

To view the commands issued to a client, first select a client:

>> select 1

Then, issue the coms command to view all the commands for the client:

(Client 1) >> coms

To view the full command and it's full output (on the coms command the output is truncated to fit the screen):

(Client 1) >> com 1

If the command output is too large for your shell, you can utilize the more command:

(Client 1) >> more 1

Workflow

After all the installations and configuration is done, the workflow of the application is the following:

  1. Client visits the home page http://localhost:5000/
  2. He makes a register request to the server
  3. The client waits for commands
  4. In the meanwhile, on the server, you execute commands using the shell.py script
  5. The client probes the server for commands, see a new one appeared, pulls it and executes it
  6. Once he's done executing, he will post back the result to the server
  7. Now, using the coms command (in the shell), we can see the output for that command

Database Handling

I have included a script that i've been using during tests, which is the db_handler.py file. It includes various function to handle and test your database. In Example:

  • List all records
  • List a specific client
  • Create the table
  • Insert a record
  • Insert a dummy record
  • Drop the table
  • Drop + Create + List the table (useful for debugging)
  • Truncate the table
  • Create a command

New: Injectable via <script> tags

Now, by visiting the page http://<yourwebsite>:<port>/js, the server will automatically generate a new injection payload that contains all the dependencies, specific URL and PORT (from the config file).

In order to comply with this feature, change the following values in the config.py file:

  • URL - the url of your website/server that contains the shell, i.e. http://yourwebsite.com
  • PORT - the port that the website/server listens to.

This method can be useful as a light-weight alternative to the Beef project.

Credits

Canop for JSON.prune

About

An interactive multi-user web JS shell

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 64.5%
  • JavaScript 33.9%
  • Other 1.6%