In0ri – Defacement detection with deep learning

in0ri

In0ri is a defacement detection system utilizing a image-classification convolutional neural network.


Introduction

When monitoring a website, In0ri will periodically take a screenshot of the website then put it through a preprocessor that will resize the image down to 250x250px and numericalize the image before passing it onto the classifier. The core of the classifier is a convolutional neural network that is trained to detect the defacement of a website. If the monitored website is indeed, defaced, In0ri will send out warnings via email to the user.

Requirement

  • Python3 (version >=3.6)
  • Docker
  • Docker-compose

Installation

Cloning the repository

git clone https://github.com/J4FSec/In0ri.git
cd In0ri

Configuring email credentials to send notifications and agent keys from

Edit the file Alert/sendEmail.py

EMAIL_SERVER = "mail.example.com"
EMAIL_ADDRESS = "foo@gmail.com"
EMAIL_PASSWORD = "$uper$ecurePa$$word"

Configure Telegram notification

Edit the file Alert/chatbot.py

CHAT_ID= 'foo' # Channel ID to send notifications to
TOKEN = 'bar' # Bot token retrieved from @BotFather

Starting In0ri

docker-compose up -d

Usage

There’s two ways to deploy and using In0ri:

  • Running off crontab by periodically visiting the url.
  • Internal agent running off the web server

First Method: URL Check

Visit the WebUI on https://<serverIP>:8080/ and click on “Register” then fill in the form and submit it.

Second Method: Internal Agent

Visit the WebUI on https://<serverIP>:8080/ and click on “Register” then fill in the form and submit it.

Click on “Create Agent” then fill in the form and check your email for the Agent’s key.

On the web server that you wants to be monitored by In0ri, download the Agent folder from Github repository

Installing the required packages for the internal Agent

python3 -m pip install watchdog
python3 -m pip install requests

Edit the file config.json in the same folder as agent

nano config.json

key is sent to your email after registering the Agent on the WebUI rootPath is the root directory of the web application that you want to be monitored exludePath are the subfolders that you wants excluded from the scans apiServer is the URL to the API server of In0ri serverIP is the IP of the API server of In0ri

{
    "id":"01",
    "key":"123123123",
    "rootPath":"/var/www/html",
    "excludePath":"",
    "apiServer":"http://<serverIP>:8088/checkdeface"
}

And run the Agent:

python3 agent.py

References

  • Model training is handled by Shu
  • Mentioned by Kitploit

Authors

In0ri is built by Echidna with the help of Cu64 and Klone.

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.