Reference implementation for "When does tail-risk optimization change a quantum gate? An equivalence boundary for mean and CVaR control"

Reference implementation accompanying the research article "When does tail-risk optimization change a quantum gate? An equivalence boundary for mean and CVaR control." This release contains the research code, analysis scripts, canonical numerical results, figure-generation material, and hardware-validation implementation used in the study of when mean-risk and tail-risk quantum control select different controls. The implementation covers the numerical experiments underlying the equivalence regime, channel-geometry and mechanism studies, nonlinear and leakage-aware simulations, colored-noise and relaxation analyses, the measurement-only predictor, predictor calibration and Ktrain controls, offset decomposition, low-kappa analysis, and the Stage 3 hardware-validation campaign. The hardware campaign is included as a validation implementation and data release. It does not constitute a positive demonstration of a CVaR advantage on hardware; the manuscript reports the observed hardware null and the associated background-floor diagnosis. The release is organized as a reference implementation rather than a dump of the complete development history. Superseded scripts, caches, smoke tests, intermediate optimization outputs, and temporary development artifacts are intentionally excluded.