The primary purpose is to automate a routine like: - unit test generation - commit messages - changelogs - commands like "how to unzip this zip in shell" or "run docker node container in interactive mode shell mount npm pack file to root ./tprompter-1.2.0.tgz" - even "ask suggest a good hackernews title for project based on the provided README" works :)

It has two modes: prompt generation and direct integration with AI API. E.g., `tprompter ask run node docker mount /home to ./ export port 123456` or `tprompter agent commit_message` will return LLM response right in the terminal (and render md!).

But `tprompter generate unit_tests --after=chatgpt` will generate a prompt with embedded source code and open ChatGPT page with the generated prompt. So, the second is a good thing if you don't want to create API keys or want to leverage the latest features of ChatGPT, but the first one will allow you not to go to a browser.

If case you are interested, I would encourage you to check the repo's README; there are a dozen examples and fancy gifs. That's the reason why I put the "toolkit" in the title; there are different ways to do things.

The project's core idea is to use the CLI possibilities, pipes, autosuggestions, fancy response rendering, integration with ChatGPT page, and so on. To be complex but provide simplicity, like Systemd.

There are predefined agents (or templates) like "commit_message" or "summary", but if you want to make your own, there is a templating engine with a catalog.

I would really appreciate any ideas or suggestions. You might contribute now if you want; there are no github actions, but I will check your code by myself at the first stage :)