An efficient way to do "reverse queries" in MongoDB for an expert system?

I am working on an "expert system" in which I would like to perform "reverse queries", that is I want to store queries and find queries that match a given object.

For example, if I have an object like this:

{
    color: 'green',
    eyes: 2,
    limbs: 4
}

I would like to return any queries that would match that object (from a collection of queries I have). Such query stored in my knowledge base would also have a parameter called conclusion that would store (you guessed it) a conclusion that this query would produce (i.e. if my object is green and has four legs, it looks like a frog).

I was thinking about writing a query creator depending on the object and storing all my "knowledge queries" in the database, however I'm not sure how efficient this would be.

Is there an existing, efficient way of doing this? Is there a document-oriented database that supports this?