You're not using them wrong at all. Part of the reason LLMs are good at writing code is that they're good at understanding code. You're just not using the full menu of capabilities -- which is totally fine.