During October 2025, Weixuan developed the ScansResource feature for the Giskard-AI/giskard-hub repository, enabling automated vulnerability scanning for AI agents within the Giskard Hub SDK. The work involved designing and implementing Python data models to represent scan results, probe outcomes, and scan attempts, ensuring robust data handling and extensibility. Weixuan integrated these models with the SDK’s API layer, providing endpoints to launch, manage, and retrieve scan results. Comprehensive documentation and targeted tests were included to support maintainability and reliability. This feature improved the hub’s security posture by facilitating automated threat assessment and enhancing risk visibility for AI agent deployments.