llm script needs the concept of a queue

[TangentFoxy]

Right now, I've been running the same batch of tests for several hours. I need to make something where I can queue up experiments and view the results later. I'm thinking a folder that is .gitignore'd in the repo holding txt files with scripts to be run. (I think to do this, llm has to be split into a library and a called script..) There would also need to be an output folder to place results in. And a way to notify or check for results.