SMS is a native macOS application for authoring Bayesian Belief Networks and running Non-Homogeneous Poisson Process Monte Carlo simulations over heterogeneous crew manifests. Built in collaboration with Dr. Eric Petersen of the APEX Aerospace Surgery Fellowship at the University of Arizona for SpaceX mission planning. The platform replaces the incorrect Bernoulli day-by-day approach used in prior tools with NHPP, correctly capturing long-tail risk distributions. Features a visual BBN canvas, FHIR R4 crew data import, Variable Elimination inference, Gibbs MCMC parameter learning, one-at-a-time and two-factor sensitivity analysis, Dirichlet uncertainty propagation, and PDF/CSV export with full provenance. 160 passing tests across three independent engine packages.