Actually for many cases the LLM already knows enough. For more obscure cases, piping in a --help output is also sometimes enough.
Actually for many cases the LLM already knows enough. For more obscure cases, piping in a --help output is also sometimes enough.