<p>This paper defends weak and local methodological continuity—that is, continuity with respect to specific shared devices rather than global methodological frameworks—across science, naturalized metaphysics, and autonomous metaphysics. The contemporary debate has conflated two distinct continuity problems: whether naturalized metaphysics can share in science’s epistemic success and whether proximity to science privileges certain metaphysical approaches over others. Additionally, continuity claims involve multiple dimensions that are not always properly distinguished. The paper makes two interconnected contributions. First, I develop a triadic framework that distinguishes these continuity problems and dimensions, providing analytical tools for practice-sensitive assessments. Second, applying this framework, I argue that thought experiments, understood as shared methodological devices, exhibit methodological continuity across all three domains. The analysis shows that thought experiments perform structurally analogous roles in each domain despite operating under different epistemic constraints and background commitments. This methodological continuity neither presupposes nor entails strong epistemic conclusions, but demonstrates that the science-metaphysics relationship resists simple dichotomies of complete continuity or categorical discontinuity. The framework suggests a productive methodological reorientation: rather than debating global continuity a priori, philosophical inquiry should examine how specific methodological devices function across domains through careful, local case studies. Thought experiments provide one such case; the approach developed here applies to other shared methodological tools.</p>