AgentVulnMitigator

A multi-agent AI security tool to detect and mitigate vulnerabilities in Agentic AI systems, aligned with OWASP Top 10 for LLM Applications.

Features

• Detects vulnerabilities like prompt injection and data leakage.
• Mitigates risks with secure protocols.
• Web dashboard for risk visualization.

Screenshots

Setup

python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
uvicorn main:app --host 0.0.0.0 --port 8002

## Features
- Detects vulnerabilities like prompt injection and data leakage.
- Mitigates risks with secure protocols.
- Web dashboard for risk visualization.
- Azure deployment with Sentinel logging.

## Setup
```bash
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
uvicorn main:app --host 0.0.0.0 --port 8002