Worked on the intelligent-machine-learning/dlrover repository to address an issue with NCCL version reporting accuracy in CI and development environments. Focused on debugging and system programming, the developer refined the shell scripting logic to exclude the VERSION_STRING entry and tightened the grep pattern, ensuring more reliable detection of the NCCL library version. This fix reduced false positives and improved the speed of triage for version-related issues, enhancing maintainability and traceability within the codebase. Utilizing C++ and shell scripting, the work contributed to more robust version parsing, supporting faster debugging and more accurate reporting for future development and maintenance efforts.