Inspiration

What it does

Checking an downloaded or an uploaded .exe or a .dll file through machine learning using Random Forest Method.(Accomplished)

We have created an application which predicts that an executed file is legitimate or malicious using a trained machine learning model with a pre-defined data set .

This ML model uses Random forest method for testing the behaviour of an application that has an executed or has a .dll file in it.

Random Forest based model helps in to solve regression and classification problems.It utilizes ensemble learning, which is a technique that combines many classifiers to provide solutions to complex problems and creates decision tree which helps us to predict the certainty of a file.

Check https://github.com/iceticshacker7/GRID_Proj/blob/main/testcase1/antivirus.py and https://github.com/iceticshacker7/GRID_Proj/blob/main/testcase1/train.py for getting the idea of how the implementation works.

In this test case we are going check the scripts through scanning and comparing the threats .Check https://github.com/iceticshacker7/GRID_Proj/blob/main/testcase2.py

How we built it

It has an automated feature where by just entering the link you can scan all files and folders of it and list them all.

It also has an upload feature where you just browse the repository in your

It can check the security of files with extensions such as .dll, .py , .exe,.cpp,.md5,.cs,.js,etc.

It predicts the vulnerability in the code and shows that the file is vulnerable or not.

It checks whether the file is legitimate or malicious.

It gives the names of the file that are dangerous.

It also what kind of activities an open source software can perform on a computer or the operating system and what kind of permissions are required to use it.

Challenges we ran into

The limitation of our project is that it cannot is not connected to the cloud we were trying to connect to scan the repository directly through the link

Second limitation is that we could not compile testcase 2 to the first one because of less time available, but will be accomplished

Third limitation is that the testcase 2 only works for github repository and can only check the files that the main folder contains.

Some scripts may have some vulnerability but are predicted legitimate Some use-cases may not work because of the altered versions of our prototype that we are unable to upload .

Accomplishments that we're proud of

What we learned

What's next for Open source software scanner

The future scope of our project is that we will add the features that in future it will get the information of the repository and what is rating of the repo , how many contributers it has , status of the admin of repository and his contributions.

We will create an automated system using cloud services that will run on cloud and the user will just have to provide the link and through cloud computing we will run the repository with our testcases then provide the feedback whether it vulnerable or not.

We will add more Testcase that will help determine the authenticity of the each type of file by using machine learning model and also add a classification technique to classify the vulnerable files and then suggest changes in it that will clear the threat and the fill will be usable.

We will create a proper interface for the project that can scan the whole repository directly .

Our team would try create a plugin that checks the repository while visiting and validate it.

In future the project will able to read and check vulnerability in all types of files.

In this application we would add more features that will check the space and time complexity of files by running it in a cloud and then return the all possible values for given testcases.

Our team will also add a feature that will check the activity the open source software will perform on the server and give a feedback for any possible threats.

Built With

Share this project:

Updates