/google_RAT

A Remote Access Tool using Google Apps Script as the proxy for command and control.

Primary LanguagePythonMIT LicenseMIT

google_RAT

Version Status

A Remote Access Tool using Google Apps Script as the proxy for command and control.

TODO

  • Support multiple masters. Any given HTTP POST/GET request from a master can fail due to another master or server having modified the Google Sheets database first for a specific client.
  • Support built-in key logging for each client type. Depending on the client type and OS platform, the ability to log user keystrokes changes.
  • Support built-in screenshot capture for each client type. Depending on the client type and OS platform, the ability to capture a screenshot image changes.
  • Support built-in "dissolve" functionality for each client type.
  • Update C2 diagram to remove number of chunks from payload format.

Dependencies

NOTE: These only apply to running master.py and test.py on your local machine. Client dependencies (if any) are listed in each client's README.

Setup

1️⃣ Deploy Google Apps Script C2 Server

NOTE: Use a private browser session for the following steps to prevent conflicts with any other Google accounts you may be currently signed into

  • Create a fake Google account (https://accounts.google.com/signup)
  • Create a new empty spreadsheet in the fake account's Google Drive (https://drive.google.com)
  • Make this new spreadsheet public and openly editable by link:
    • File > Share > Get Link > Change > Anyone with the link > Viewer > Editor
  • Paste the new spreadsheet's link into the SPREADSHEET_URL variable in server.js and define a secret value for MASTER_KEY.
    • NOTE: Remove ?usp=sharing at the end of the SPREADSHEET_URL. The URL should end in /edit only.
  • Visit Google App Scripts (https://www.google.com/script/start/) and make a new project under your new Google account:
    • Start Scripting > New Project
  • Paste your now formatted code from server.js and save the project
  • Publish the project (following steps from Google documentation):
    • Deploy (top right corner) > New Deployment > Web App (as the deployment type)
      • Fill in the description field with something
      • Make sure the app is executed as Me
      • Make sure Anyone can access the app
      • Click Deploy
      • Click Authorize Access > Your fake account > Advanced > Go to ... (unsafe) > Allow
        • NOTE: If you do not see this step, make sure you are using a private browser session
    • Save the application URL (it should end in /exec). This is what the clients and master will connect to.

2️⃣ Test Server Connection

  • Run ./client/test.py in order to test your server URL connection and MASTER_KEY:
    • NOTE: Running this test will leave an empty inactive client in the Google Sheets database. Simply delete that row to remove this inactive client.

test

3️⃣ Select Clients

  • Select your client and add the Google Apps Server URL from step 1 into the correct payload variable for your client's type as defined in the client's README

4️⃣ Run Master

  • Run the master to interact with clients:

master

Command and Control Protocol Notes

NOTE: diagrams made with https://draw.io

  • Transaction Flow:

architecture

  • Client State Transition Diagram:

state

  • Example server transaction between a master and client in Google Sheets:

server

  • General Notes:

    • This design allows for multiple servers to be ran simultaneously against the same backend Google Sheets "database" for client redundancy and availability.

    • All master requests to the server must present a unique key in order for their request to be processed. This key is hardcoded into each server's JavaScript with the MASTER_KEY variable.

    • Each payload is base64 encoded except for the the command type. This is seperated by the | character as the delimiter in the payload.

Limitations

  • All data sent to/from the server is chunked into 50000 (50 KB) chunks. This is because Google Sheets currently has a single cell size limitation of 50000 characters:

  • Google applies daily quotas and limitations for execution of its services. Getting around these limitations is as simple as creating other duplicate copies of the same server.js code for more servers in your design. Each client is able to cycle through multiple servers for loadballancing.