I told it "Here is a shell prompt, try running a command..." then copied its command into the shell, copied the output back into ChatGPT, with the prompt "Here's the output, try another" and so on.

I did the same yesterday. don't forget to add the following to the prompt: "for dangerous commands, such as `rm` emit a "danger!" warning.

if such a warning occurs, I insert a confirmation dialogue.

Next: parse and compile the code and language in

```language blocks ```

If you are interested in doing this in an automated fashion, see my GitHub repo. It uses telegram as a frontend for gpt executing code on the backend.

https://github.com/thornewolf/gpt-3-execution

I haven't tried this with recent versions, but this was great fun early on:

https://github.com/f/awesome-chatgpt-prompts#act-as-an-ai-tr...