Worked on the alibaba/loongsuite-python-agent repository to enhance observability by addressing two key bugs related to OpenTelemetry span naming. Focused on improving trace readability and test reliability, the developer corrected the dynamic inclusion of function names in ToolCall spans within the Agno agent and updated test assertions to match new naming conventions. Using Python and leveraging skills in debugging, observability, and testing, the changes improved instrumentation reliability and reduced false positives in downstream metrics. The work demonstrated attention to detail in aligning code and tests, ensuring that OpenTelemetry traces accurately reflect agent behavior for more effective monitoring and analysis.