Unlike other MS applications (like Notepad), the command prompt is no configured for full dictation. It can only understand individual letters and symbols through MS Speech recognition. Although that may be sufficient for short MS commands, it isn't satisfactory for console applications using the CLI where full dictation might be more useful than simply typing data in.
One way I suppose this could be implemented is by using Microsoft's SAPI to convert speech into text and then to read the output from a text file to make it appear that the CLI has understood full speech. But that seems very long-winded. Is there a quicker way to reconfigure the command prompt's parameters to enable full dictation of entire words and not just characters?
I'd normally just comment on something like this instead of list an answer. Stack Overflow is saying I need a street cred of 50 to do that... for whatever reason.
You'd probably have to create a custom solution, which would require emulating the command line interface and executing programs via shell commands and creating your own grammar specific for common things performed in the terminal/command line interface. Then you'd have to create dynamic grammars on the fly that would be relevant to your user (such as creating grammar for each file and directory available to him or her.)
Needless to say, this in itself would be a huge undertaking. Short of that I don't believe there is a better solution than what is currently available in a shared recognizer context.