Description
I instruct Claude Code to always use rtk read, rtk grep etc instead of the built-in read tool, to make sure it doesn't simply default to using its own built-in tools and thus not save any tokens. The problem is that the Claude Code built-in read tool supports reading a specific range of line numbers, so it tries to e.g. run rtk read <file> <start_line> <end_line>, which first returns the full file contents and then throws two errors. It would be nice if rtk read would support reading a range of line numbers, so that I don't need to explcitly instruct the LLM to use its built-in reading tool if and only if it reads a specific range.
I'm only a newcomer to rtk.ai so if a clean solution already exists, please let me know and I will close this issue!