Worked on improving the reliability of the Search Agent workflow in the aws-samples/generative-ai-use-cases-jp repository by addressing a configuration issue that blocked user input. Focused on correcting the ActionGroupName from 'UserInput' to 'UserInputAction', which enabled proper user interaction and ensured alignment with existing snapshot tests. This change enhanced end-to-end reliability and maintained test coverage for user-interaction scenarios. Utilized AWS CDK and TypeScript to implement the fix, demonstrating attention to configuration integrity across environments. The work concentrated on bug resolution rather than feature development, reflecting a targeted approach to maintaining robust and testable cloud-based application workflows.