Metacognition seems pretty universally positive. I'm wary of viewing it as such. Aside from the obvious criticisms like "you can't learn to ride a bicycle by thinking about and writing a 200 page treatise on riding bicycles" I'd like to know some concrete places where metacognition should be discouraged in mathematics education. Wittgenstein, for example, may have argued that we acquire the ability to do mathematics by implicitly learning what not to think about or reach for in a very unconscious way akin to natural language use. Even if one views Wittgenstein's views on mathematical practice as fringe, one might note that the typical (or perhaps I should say "ideal" in the sense of Reuben and Hersh) mathematician's aversion to philosophizing about mathematics reflects a general attitude against metacognition in mathematical practice…perhaps fearing one's never getting down to business because of being lured by the sirens of philosophy. Here, though I would like more concrete criticisms in the everyday practice of teaching.
Being an odd mathematician in terms of my attitude toward philosophy, I would love it if metacognition were universally good for students, as then we could just have them write in order to build in some continuity of concepts. I pause, though, when presented with a formalist critique akin to a Wittgenstein-like position above.