Current immunotherapy aims to reprogramme the tumor microenvironment to improve immune recruitment and tumoricidal activity, but its success ultimately depends on antigen recognition. Here, we propose introducing viral antigens into tumor cells to redirect pre-existing virus-specific memory T cells for cytotoxicity, leveraging the global prevalence of SARS‑CoV‑2 vaccination and infection to ask whether established anti–SARS‑CoV‑2 immunity can be repurposed for cancer therapy. To test this concept in vitro, patient-derived head and neck squamous cell carcinoma (HNSCC) cells were transfected with a commercially available COVID‑19 mRNA vaccine and co-cultured with spike-enriched T cells isolated from partially HLA-matched vaccinated donor PBMCs, resulting in preferential killing of spike-expressing tumor cells versus lipofectamine-only controls, accompanied by increased frequencies of IFNγ+CD8+ T cells; importantly, cytotoxicity was abrogated by HLA class I blockade. We then validated this strategy in vivo using NOD-scid IL2Rγnull (NSG) mice bearing subcutaneous patient-derived HNSCC, treated with repeated intravenous infusions of spike-enriched T cells together with intratumoral COVID‑19 mRNA dosing, resulting in delayed growth relative to an irrelevant vaccine control. Finally, in a vaccinated syngeneic HNSCC model (MOC1 cells), intratumoral COVID‑19 mRNA with or without anti‑PD‑1 delayed tumor growth, and combination therapy increased intratumoral CD8+ infiltration; deep tumor phenotyping further revealed distinct waves of spike-targeting and tumor-targeting T-cell infiltration, suggesting antigen spreading. Together, these results support harnessing widely prevalent cellular immunity to SARS‑CoV‑2 as a practical route to redirect cytotoxic T-cell activity against tumor.