We now stand firmly in the era of rocky exoplanet detection, yet the empirical characterization of these most intriguing planets is extremely challenging. Transit plus radial velocity information can yield planet mass and radius, and hence planet density, but the bulk composition remains degenerate and model−dependent. The abundances of a handful of exoplanet atmospheres can be estimated from transit data, or observed directly via spectroscopy, but probing only the tenuous outer layers of those planets. Fortunately, there is overwhelming evidence that polluted white dwarfs can yield highly accurate information on the chemical structure of rocky minor planets, the building blocks or fragments of solid exoplanets. The white dwarf distills the planetesimal debris, and provides powerful insight into the mass and composition of the parent body. This 'archaeological' method provides empirical data on the assembly and chemistry of exo-terrestrial planets that is unavailable for any planetary system orbiting a main−sequence star. In the Solar System, the asteroids (or minor planets) are leftover building blocks of the terrestrial planets, and we obtain their compositions -- and hence that of the terrestrial planets -- by studying meteorites. Similarly, one can infer the composition of exo-terrestrial planets by studying tidally destroyed and accreted asteroids at polluted white dwarfs. I will present ongoing, state of the art results using this technique, including evidence for distinctly terrestrial-like compositions, chemical differentiation, and the detection of water-rich planetesimals that may represent the building blocks of habitable exoplanets.