Worked on the strands-agents/sdk-python repository to improve the reliability of telemetry span ordering within agent development workflows. Addressed a bug where the system prompt was not included before input messages, which previously affected trace accuracy and the consistency of agent interactions. Applied a targeted fix in Python, focusing on the structured_output_span logic, and expanded the test suite to validate correct span sequencing and reduce the risk of regressions. Emphasized maintainability by providing clear commit documentation and ensuring that telemetry traces now accurately reflect the intended message flow. Utilized skills in Python, telemetry, and testing throughout the development process.