Show HN: Omega Walls–open-source stateful runtime defense for RAG and AI agents

Category: security

Tags: security, ai-safety, rag, runtime-defense, prompt-injection

Score: 6.5/10 (Innovation: 7, Technical: 6, Documentation: 8, Utility: 5)

Omega Walls is an open-source stateful runtime defense system for RAG and AI agents that protects against distributed, multi-step prompt injection attacks. It's interesting because it moves beyond single-turn classifiers by accumulating risk across sessions and enforcing deterministic actions like blocking, freezing, or quarantining before dangerous tool execution.

Target audience: ai engineers, backend devs, security engineers

Repository: https://synqra.tech/omega-walls · Python · Apache-2.0 · 2 stars

View on Hacker News