Covalent inhibitors of oncoprotein KRAS have initial efficacy, but responses lack durability. Covalently modified oncoproteins are presented as MHC-restricted hapten-peptides (p*MHC) on the cancer cell surface, enabling combination of targeted therapy with immunotherapy to overcome drug resistance. Building on indirect evidence of KRASG12C-derived p*MHCs, we use immunopeptidomics to identify and directly quantify these synthetic neoantigens. To address challenges by their low copy number, we develop AETX-R114, a T cell engaging bispecific antibody with picomolar affinity for MHC-restricted sotorasib-modified KRASG12C peptides presented by three HLA-A3 supertype alleles. AETX-R114 dramatically increases the half-life and thereby the number of presented p*MHCs, enabling selective and potent killing of resistant cancer cells both in vitro and in vivo. To broaden the therapeutic potential of creating and targeting synthetic neoantigens, we further develop AETX-R302, which recognizes divarasib-modified KRASG12C peptides presented on alleles from the HLA-A2 and A3 supertypes. Cryo-EM structure determination reveals the molecular basis for breaking HLA supertype restriction. Collectively, our study illustrates how engineered antibodies can transform synthetic neoantigens into actionable cancer immunotherapy targets.