I built Lyzr's production Responsible AI engine, the guardrail layer that sits between enterprise AI agents and the real world. It provides PII detection and redaction, toxicity metering, NSFW, gibberish, and prompt-injection detection, and topic controls that keep agents inside their intended scope.
Policy enforcement is pluggable: the engine integrates OPA, Cedar, AWS Bedrock guardrails, and Google Vertex guardrails, so enterprises can bring the policy stack they already trust. I also integrated the engine into ComputerAgent's open-source and enterprise editions as a chained guardrail redaction proxy with policy-bound agent runs.