I am developing a skill that performs a knowledge base search. The utterances are of the form "What ..." where the anchor phrase can be who, what, where, when, why or how, or a contraction of the form "What's ..." Each anchor phrase causes a specific intent to be triggered to process the single instance of slot data which is declared as AMAZON.SearchQuery. This all works well.
When the utterance does not contain one of the anchor phrases, e.g., "I want to order a pizza,"I am relying on the AMAZON.FallbackIntent to trigger so I can handle the out-of-domain utterance. This does not appear to work at all.
I never see the Fallback intent being triggered. Sometimes I get an audio tone while other times I seem to be triggering some other skill and receiving responses that I have not coded into my lambda function. Since this skill is targeted at the under-13 age group, randomly selected skills/responses are simply unacceptable.
I have two questions: (1) does the fallback intent actually work as documented (https://developer.amazon.com/docs/custom-skills/standard-built-in-intents.html#fallback); and, (2) is there any way to restrict the triggered intents to only those within my skill?