@elegant-branch-54423 - I'm inclined to agree - but there are a few other things I think we need too - regarding my PR above - I have been trying to address: • The agent builds scripts which have incorrect expectations about the environment in which they run. (When I asked it why it built an automation, it essentially responded with

What is a sandbox anyway, and where am I?

) • The agent wrote scripts which have incorrect expectations about the environment variables (The examples referenced variables that are not available to the scripts) • The agent wrote scripts which it did not check and contained invalid python