Build an open-source assessment tool that measures 'cognitive dependency' by tracking how often users rely on LLMs versus manual problem solving in a coding environment. This could provide metrics on user skill stagnation.