Show HN: Runtime security for AI agents(injection,tool abuse, data exfiltration)

Category: security

Tags: ai-security, llm, prompt-injection, policy-engine, runtime-security

Score: 6.0/10 (Innovation: 6, Technical: 6, Documentation: 4, Utility: 6)

AI-SPM is an open-source runtime security layer for LLM pipelines that detects prompt injection, enforces structured tool calls via policy engine (OPA), prevents data leakage, and streams activity for audit. It aims to treat the LLM as untrusted and enforce security externally, addressing a gap in AI security similar to CSPM for cloud.

Target audience: security engineers, ML engineers, backend developers

View on Hacker News