If you want to mix FCI and AG, you can do so and it is supported. Below is a hypothetical configuration.
Failover Cluster Instance (Node A, Node B), addressed as Server 1. All clients connect to Server 1. Only one Node is live at a time (i.e. only one node has control of the storage that has the databases on it.
Node C lives in your Disaster Recovery center. You don't want to go through the expense of making your SAN geographically dispersed so you can't just add it to the Failover Cluster Instance. So you setup an Availability Group. That availability group is on Node C <--> Server 1 (A/B).
Given my setup and my businesses existing tolerances for data loss, I would setup the AG for asynchronous commit and manual failover.
Direct Answers to questions
1 - The FCI can be either the primary or the secondary member of an availability group. To the availability group, it's just a server. Point the AG at the FCI cluster name though.
2 - Both nodes will participate (one at a time) in the AG. FCI is like two brains sharing a body. Only one can control the body at a time, so for all intents and purposes only one brain exists at a time. An AG is two complete setups, one shadowing another ready to step in at any time.