Find Evil Agent

• 

  Running

Inspiration

as a beginner in cybersecurity, i was inspired by the growing threat of cyber attacks that happen in minutes. i wanted to build something that could respond automatically. a defender that never sleeps.

What it does

find evil agent is an autonomous cybersecurity incident response agent that:

• analyzes system logs and detects suspicious activities with exact line-number tracking
• uses AI (Groq Llama 3.1) to generate structure investigative narratives -validates its own findings and self corrects if the AI response is inaccurate -export a complate JSON report for audit trail.

How we built it

-python 3.12 on ubuntu WSL2 (windows) -Groq API with Llama 3.1 model for AI Reasoning -Custom log analysis tool with keyword detection -sel correction loop with up to 3 retry attempts -JSON report export for accuracy documentation

Challenges we ran into

-setting up WSL2 and resolving DNS connectivity issues -managing API rate limits during development -ensuring the AI validation logic was accurate and consistent

Accomplishments that we're proud of

-how autonomous AI agents work in cybersecurity -how to build self correcting AI systems -log analysis and incident response fundamentals

What we learned

-support for more log formats (windows event logs, syslog) -real time log monitoring -web dashboard for visualization

What's next for Find Evil Agent

Built With

• groq
• llama-3.1
• python
• python-dotenv
• ubuntu-wsl2