[Sorry for grammar... english is not my first language]
There's something I don't understand about slot parsing in a test skill I'm writing.
I have a custom slot type defined as a list of italian colors (argento,oro,bianco,grigio,blu,verde,giallo,arancio,rosso,marrone,nero)
I have an intent called QuantoValeIntent, with must be invoked with "Quantro vale {coloreA} {coloreB} {coloreC}" or "il valore di {coloreA} {coloreB} {coloreC}"
The intent slots are all required, and in the validation settings all have "Value within slot type's slot values".
Now... if I evaluate the model with "il valore di rosso nero giallo" (or every color beside blu) the selected intent slots are filled with the colors
BUT... if the color "blu" is involved, things get interesting.
A call with "il valore di blu nero giallo" the "blu" is totally ignored!
Best of all, a call with "il valore di blu blu blu" returns a completely bad parsing
It seems that Alexa doesn't like the blue color... and I don't know why. It's not the word "blu" too short, because the "oro" one works flawlessly.
Where am I doing wrong?