Post-LLM validation. We're currently working on this at https://github.com/guardrails-ai/guardrails