Hi HN! After a year of AI red teaming, I published our internal security baseline as an open standard.

The Baseline covers: - Pre-deployment (threat modeling, prompt injection testing) - CI/CD integration (automated security gates) - Runtime protection (I/O filtering, rate limiting) - OWASP LLM Top 10 mapping

Tools to implement it (free to try): - AgentAudit: https://app.xsourcesec.com - 650+ attack vectors - BreachLab: https://breachlab.xsourcesec.com - Gamified prompt injection training

Happy to answer questions about LLM security!