This has nothing to do with PNaCL or Dart, so why beat that strawman? This is about standards committee stuff being pushed out before finalization not behind a flag, the kind of stuff that incited the big brouhaha over prefixed APIs not too long ago. Commitment to implement is not the same as "won't change before final release", so what happens if it gets significant uptake, and the final draft actually tweaks something important?
As far as I know, the way ES6 works is that there is consent for everyone to implement features that have been approved in the "face to face" meetings. That is not the same way that the CSS working group works, however, because CSS is now modularized as opposed to monolithic.
You're trying to create controversy where there is none. Nobody on TC39 objects to SpiderMonkey implementing specific features with broad consensus before all of ES6 has been finalized.
What's good for the goose is good for the gander.