Agent Sentry

Inspiration

As AI agents gain access to sensitive data and business systems, we wanted to solve a simple problem: How do organizations know what their AI agents can access, control what they do, and investigate when something goes wrong?

What it does

Agent Sentry is an AI agent security and compliance platform. It helps organizations monitor agents, manage and revoke data access, enforce permissions, investigate incidents using PRISM, and stay aligned with changing regulations.

How we built it

We designed a centralized dashboard with agent inventory, access management, and incident investigation. Our architecture uses FastAPI, Next.js, PostgreSQL, Neo4j, and policy-based enforcement, with planned PRISM integration for tracing agent activity.

Challenges we ran into

The biggest challenges were defining precise access permissions, connecting agent activity to security incidents, understanding the impact of regulatory changes, and keeping the dashboard simple.

Accomplishments that we're proud of

esigning a unified platform that brings agent visibility, access control, incident investigation, and regulatory intelligence together in one place.

What we learned

We learned that AI security isn't just about monitoring agents. It's about controlling what they're allowed to do, enforcing those boundaries, and being able to explain every important action.

What's next for Agent Sentry

We're working toward a functional prototype with real-time permission enforcement, access revocation, PRISM-powered investigations, and AI-assisted regulatory change analysis.

Built With

• claude
• cloudflare
• fastapi
• niccron
• prism
• python
• react
• typescript