Go over the specs with designers and rank elements to determine what needs to be more or less prominent and do more research on where we disagree. I often involve a dev also to make sure there aren't technical constraints.
Typically, we would end up doing quick napkin drawings. What fits on paper does not necessarily mean it will
fit on screen.
Create wireframes in 'min spec' screen size knowing that it might feel cramped and that most of the users use regular size and is going to feel great.
Build a quick prototype using the wireframes. During this step, if I see potential issues, I would bring it up with team and address it immediately, and not wait until I present it to them to reduce revision cycles
Let people play with the prototype.
Go back to step 02 as necessary
Skinning phase. Until now, everything should be in grey scale. Depending on the stage of the project this step will be different.

If no style is established yet:
Present several skin variations of a few screens

If a style is established:
Skin at least the major screens

If this is a new feature of an existing product:
Build it in the product. It is often more efficient than to create high fidelity mocks.