eg. given a cli specified as

?:input Hello there what can I do for you?
if input == "build"
  Ok, I'll build the app for you, hold on a sec -> ./build.sh
elseif input == "deploy"
  Ok, I'll deploy the app for you, hold on a sec -> ./deploy.sh
else
  Sorry, that's not a valid command, by then -> exit

let's say we compile this to an executable called foo.

Can I simply do foo build if i want to build, instead of first running foo, waiting for the prompt and then inputting build?

Could not the same functionality be provided using an internal DSL?

Meaning without inventing a new syntax and language?