GQMI: MedAI Trust-Graph (v0.2.3) — Audit-first haemodynamic design test and dynamic clinical consistency

What this is This release ships an audit-first, NO OVERCLAIM governance framework for Medical AI, integrating: M1: a trust-graph of the medical AI pipeline (fixed node IDs, trust-injection points, responsibility anchors), M2: a haemodynamic audit ledger that enforces reproducibility as a contract (same inputs + same BC + same solver config -> same outputs), M3: a dynamic clinical consistency layer (temporal coherence, causal plausibility, stability under drift/perturbation). What this is NOT This work is not clinical advice. It does not provide treatment recommendations, patient-specific guidance, or outcome claims. It is a governance and reproducibility protocol for auditable haemodynamic design tests (e.g., bypass/stent geometry changes and their effect on flow/pressure/WSS) under locked boundary conditions and solver configuration. Files Main manuscript: PDF + DOCX. Figures: Fig.1–3 (PNG). Supplementary Information: PDF + DOCX. INSTALL_ME: plain-text packing slip (how to read / what is locked / font policy). Version note v0.2.3 extends v0.1 by adding M2 (haemodynamic audit ledger) and M3 (dynamic clinical consistency), plus Figures 1–3 and SI.