Show HN: Lelu – catch AI agents when they're manipulated at runtime

Category: security

Tags: ai-security, authorization, prompt-injection, agent-monitoring, go

Score: 7.5/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 7)

Lelu is an authorization engine for AI agents that prevents manipulation like prompt injection and anomalous behavior during runtime. It provides a unique layered pipeline for agent action authorization, combining traditional policy evaluation with modern AI-specific threat detection.

Target audience: backend devs, ai engineers, devops

Repository: https://github.com/lelu-auth/lelu · TypeScript · MIT · 32 stars

View on Hacker News