LogSpear

• 

  EV Charging Devices

• 

  High Volume Devices

• 

  Network Graph

Inspiration

This forensic tool was designed for SOC Analysts who want to find suspcious IoT devices that have been used in Botnets through their log scans. The faster the SoC can find the infected IoT devices such as CCTV's and EV charging stations on the network, they can prevent future devices from becoming infected.

What it does

An LLM-powered incident summarizer. Users upload their log files to the chatbot interface to find the suspicious IP, Machine Information, Country of where the botnet is from, and a visualization in a graph of the botnet like devices connected to it.

How we built it

We built it using an AI Agent with Chainlit, OpenAI API, and LangChain.

Challenges we ran into

The main challenge was setting up Chainlit web interface with langchain and coming up with the correct log files to analyze.

Accomplishments that we're proud of

I am proud of helping Cybersecurity Analysts find these networks before they form and neutralize the threat before future devices get infected.

What we learned

I learned how to use chainlit and langchain to help cybersecurity SoC's.

Tech Stack

LLM Framework - Langchain Interface - Chainlit Language - Python Data Parsing - Langchain JSON/CSV Loaders Deployment - Vercel

What's next for LogSpear

Integrating local, proprietary models, so that the data never leaves the network.

Built With

• chainlit
• langchain
• openai