I'm not intending to denigrate your tool. I'm just commenting on the "look at how we go full circle" aspect.

I will point out that the shell script way uses less resources than an LLM making a tool call. But I understand that these scenarios are not necessarily meant for the same user.