Two steps prompt generation for LLM in command guessing

[Silin Meng]

Step One: Have it determine which type of code the user is executing (intent detection).
Step Two: Based on the result of Step One, have it generate the next possible command (command generation).