Cyber Shield Squad

Inspiration

With AI agents becoming central to industries like finance, healthcare, and enterprise, we saw a critical gap: security and trust. Prompt injection, data leakage, and malicious agent behavior are rising threats. Our inspiration was to build a guardian layer that protects AI interactions in real time, ensuring safe adoption and transparency.

What it does

AI Agent Security Guardian acts as a protective shield for AI agents.

Monitors agent interactions continuously.
Detects and blocks malicious prompts or unsafe actions.
Provides a dashboard with alerts and insights.
Ensures compliance, trust, and reliability for enterprises.

How we built it

Frontend: React + Tailwind for a dynamic UI.
Backend: .NET + Spring Boot for robust APIs.
AI Layer: Python (NLP + anomaly detection).
Deployment: Dockerized and hosted on Render.
Collaboration: GitHub repo for version control and teamwork.

Architecture flow:
User → AI Agent → Security Guardian → Safe Output

Challenges we ran into

Detecting subtle malicious prompts without blocking legitimate queries.
Maintaining real‑time monitoring without performance bottlenecks.
Balancing security vs. usability for end users.
Deploying seamlessly across multiple environments with Docker + cloud hosting.

Accomplishments that we're proud of

Built a working demo with real‑time defense.
Designed a scalable architecture ready for enterprise use.
Created a dashboard that makes security transparent and actionable.
Collaborated effectively as a team under hackathon pressure.

What we learned

How to integrate security monitoring into AI workflows.
Importance of explainability in AI defense systems.
Technical depth in combining multiple stacks: React, .NET, Spring Boot, Python, Docker.
The value of teamwork and rapid iteration in hackathons.