Maintaining a sufficient short-circuit ratio (SCR) is essential for the secure operation of high-voltage direct-current (HVDC) sending-end grids, particularly for mitigating transient overvoltages following DC-blocking faults. However, incorporating SCR requirements into unit commitment (UC) remains computationally challenging because rigorous SCR evaluation requires the inversion of a decision-dependent network admittance matrix, resulting in an implicit coupling between UC decisions and grid strength. To address this challenge, this paper proposes a novel SCR-constrained UC framework with a decoupled offline-mapping-to-online-optimization architecture, explicitly embedding grid strength requirements into system scheduling. Specifically, a combinatorial encoding technique is developed to establish an exact offline mapping between UC schemes and their corresponding short-circuit capacities (SCCs), thereby reformulating the implicit nonlinear SCR constraints into exact mixed-integer linear constraints. Numerical results demonstrate the effectiveness of the proposed framework and confirm that it preserves the standard MILP formulation of UC, enabling seamless integration into existing industrial UC frameworks.