Computational supplement for Sharp continuous-time bounds on rehabilitation exercise exposure from sparse kinematic measurements

This B-specific archive supports "Sharp continuous-time bounds on rehabilitation exercise exposure from sparse kinematic measurements". It includes 360 synthetic trajectories, six recorded B result files, B-only source code and 27 implementation tests, and author-refined Figures 1–4 in five formats. It contains no patient records, credentials, A/C/D code or A/C/D outputs. Reproduction verified the outputs within stated floating-point tolerances; 2,916 exhaustive path checks had zero failures. Implementation checks do not establish clinical effectiveness or patient safety. Supplied refined artwork includes presentation edits and is not byte-identical to regenerated plots. Code is MIT; data, figures and documentation are CC BY 4.0, with scope identified inside the archive. ChatGPT (OpenAI) substantially assisted mathematical development, programming, numerical analysis, plotting code and drafting; Xiaoqi Wang is responsible for the final scientific claims. This is a computational supplement, not a manuscript preprint or journal acceptance.