[web-animations] Drop reference to composite modes when using implicit keyframes #2914

In "the effect value of a keyframe effect", the spec currently defines the behavior when there is a missing 0% or 100% keyframe in with steps such as the following

If there is no keyframe in property-specific keyframes with a computed keyframe offset of 0, create a new keyframe with a computed keyframe offset of 0, a property value set to the neutral value for composition, and a composite operation of add, and prepend it to the beginning of property-specific keyframes.

This suggests that one needs to use addition in this case, but in fact the behavior is simply to use the underlying value (the idea of adding a null value to the underlying value--specifically a null value that is specifically defined as producing the underlying value when added to it--is a roundabout way of saying, "Just use the underlying value" that was added back when we did interpolation before composition).

We should just drop this complexity and say to use the underlying value in that case.