Enhancement idea: implement a self help

Hi!

Thanks for your amazing tool!

By the way: since it has so many features - how about implementing a kind of self help where you can ask the tool a question about it's usage, by feeding the help texts it has stored anyway and the users question to an LLM?

Not quite sure how that would best be done for llm, but I just implemented that for my AIGenPipeline https://aigenpipeline.stoerr.net/ in a couple of minutes by embedding the raw markdown files for the website into the command line client and feeding those to the backend LLM the tool uses (https://github.com/stoerr/AIGenPipeline/pull/24) : e.g.

    aigenpipeline --helpai "Make a command line where I can overwrite the file out.txt by executing prompt.txt on input.txt , no version comment."

gives you

    To overwrite the file `out.txt` by executing `prompt.txt` on `input.txt` without adding a version comment, you can use the following command line:
    
    ```
    aigenpipeline -wo -o out.txt -p prompt.txt input.txt
    ```

The issue does not name specific files or tests. Start by reviewing the command-line help entry point and the help texts the tool already stores, then determine how a user question would reach the configured LLM. Done would require an agreed design for feeding that help content to the model and a documented command-line interaction.