Command line output parser

[leftmove]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

Create an output parser for commands in the terminal. Parsers already exist for JSON and other formats, but one would be useful for terminal commands and arguments.

Ideally, you write the prompt, and get back a single string with commands separated by new lines. Right now, you could probably get similar functionality by asking for a JSON array of strings, but a parser specifically for new line separated strings would be simpler.

Ensuring that the response is compatible with the command line would allow for more efficient and safe LLM use in the terminal - especially when creating abstract, multi-line CLI scripts.

Motivation

I'm considering the idea of building an all-in-one LLM client for the terminal, and a feature where you can quickly prompt LLMs for scripts would be useful.
With an output parser, the time between prompting and running (and the complexity overall) would be greatly decreased.

Proposal (If applicable)

I would be willing to code this feature for the Python and JavaScript libraries, if given the go-ahead.