A wildcard is a placeholder for any and all input, it is not possible to specify which.
The speech recognition engine does tend to recognize numbers relatively well, in my experience, so you could use wildcard or dictation for this.
For numbers between -1 and 10, which are written out (E.G. "one" instead of "1"), you'll want to use the "{TXTWORDTONUM:}" token, which can then be passed into the "{TXTNUM:}" token to filter out anything else that may have been recognized unintentionally
E.G.
{TXTNUM:"{TXTWORDTONUM:"{CMD}"}"}