Developed generic guardrail policy support for hosted agents in the Azure/azure-dev repository, enabling users to attach Responsible AI guardrails through agent.yaml configuration. The work involved evolving the manifest and model to support a generic, typed policies list, implementing validation logic to ensure non-empty policy names, and mapping policy identifiers to the Foundry data plane. Leveraging Go for backend development and YAML for configuration, the developer introduced schema changes and wired new policy structures into container and code-deploy builders. Comprehensive unit tests and data fixtures were added, enhancing governance, extensibility for future policy types, and safer hosted agent deployments.