A Dragonfly engine backend for macOS Voice Control would be nice to have, although there are no plans that I'm aware of to implement that. I am also unable to find an API in Apple's developer documentation that's explicitly for interacting with the new Voice Control feature. I guess I could have just missed it though?
Rather than implementing a new engine, it should be possible to get the
Kaldi engine working on macOS and utilise
Aenea's mac implementation to get AppContexts working properly. This way, you could use your Dragonfly grammars like normal and swap to Voice Control for dictating prose and other tasks where it's more suitable.