By combining cosmological simulations with Frontier Fields project lens models we find that, in the most optimistic case, galaxies as faint as $m ≈ 33 - 34$ (AB magnitude at $1.6 \rm μ m$) can be detected in the Frontier Fields. Such faint galaxies are hosted by dark matter halos of mass $∼10^9 M_⊙$ and dominate the ionizing photon budget over currently observed bright galaxies, thus allowing for the first time the investigation of the dominant reionization sources. In addition, the observed number of these galaxies can be used to constrain the role of feedback in suppressing star formation in small halos: for example, if galaxy formation is suppressed in halos with circular velocity $v_c