Worked on stabilizing the autonomous subsystem for the mcgill-robotics/AUV-2025 repository, focusing on restoring system reliability and ensuring safe mission execution. Addressed a critical bug by rolling back sensors, planners, and controls to a previously stable configuration, which improved the predictability of autonomous runs and reduced mission risk. The technical approach involved repository hygiene, including correcting file renames, fixing spelling inconsistencies, and removing deprecated files to maintain a clean codebase. Utilized C++, Python, and ROS to implement and validate the rollback, emphasizing system integration and software engineering best practices to prevent regressions and support ongoing development stability.