The standard cosmological model (GR+ΛCDM) enjoys remarkable empirical success. Yet a systematic question has received little attention in the philosophical and foundational literature: are the applicability conditions for deploying general relativity at cosmological scales actually satisfied? This paper develops and applies an operationalised validation-licensing protocol to answer that question. The protocol evaluates six applicability conditions (A1–A6), identifies diagnostic signatures, and produces a reproducible verdict—PASS, OPEN, or Limit of Applicability—together with an explicit audit trail.Three principal results emerge. First, a regime-dependent applicability profile shows that GR transitions from PASS in the solar-system regime (R1) to OPEN in the cosmological regime (R3), with the transition correlating with the onset of gravitational anomalies near the MOND acceleration scale a₀. Second, a constraint-disclosure inventory identifies twelve typically undeclared modelling choices that constitutively carry the empirical fit of ΛCDM, demonstrating that this success is a configuration achievement rather than a model-independent result. Third, a diagnostic inheritance structure shows that standard necessity claims for dark matter and dark energy inherit the OPEN status of the GR-applicability presupposition on which they depend.The paper concludes with reformulation directives specifying what observational evidence would convert the diagnostic status from OPEN to PASS, and demonstrates that the protocol is diagnostic and model-neutral: it does not advocate modified gravity, does not propose an alternative cosmology, and does not predict observational outcomes.Additional Description:Preprint. Submitted to Studies in History and Philosophy of Science, Virtual Special Issue ‘Modeling the Cosmos: Frontiers in Philosophy of Astrophysics and Cosmology’ (Guest Editors: Silvia De Bianchi, Marco Forgione, Federico Viglione). Part of the research programme ‘Applicability Diagnostics in Fundamental Physics’ (Zierhut & Schwarz).
Zierhut et al. (2026) studied this question.