Read from stdin like llm-jq does

[deepakjois]

I see llm-jq has this code which read the example from stdin and appends it to the prompt

https://github.com/simonw/llm-jq/blob/beaada20d11dfc34618394c633bfa128a7c00a94/llm_jq.py#L42-L46

        is_pipe = not sys.stdin.isatty()
        if is_pipe:
            example = sys.stdin.buffer.read(length)
        else:
            example = ""

Does it make sense for llm-cmd to consume the stdin when constructing the prompt. I am thinking of a case where i want llm-cmd to generate a command with context of an output from another command (for e.g. ffprobe)

If this is of interest, i can try to do it myself and submit a PR