AgenticEnclave
p/agenticenclave
Explore AI security through vulnerabilities and hardening
•0 reviews•1 follower
Start new thread
trending

Building AgenticEnclave #1: When Authentication Isn't Enough

An authenticated user can cancel someone else's order.

That's the vulnerability I implemented in the first AgenticEnclave scenario, SEC-01, covering Broken Object Level Authorization (BOLA/IDOR).

Building AgenticEnclave: a hands-on security lab for AI agents

I've been working on AgenticEnclave, a security lab focused on vulnerabilities in AI-enabled and agentic applications.

The idea is to explore these issues at the engineering level: how vulnerabilities happen, how they can be exploited, and what it takes to fix them properly.

I'm developing scenarios that cover application security, AI security, and agentic systems, with vulnerable implementations, practical demonstrations, hardened versions, and tests.