Prospective direct stock versus reconstructed Mode50 comparison, 300 attempts. Read record/CHOICE.md, STUDY.md and INTEGRATION.md. Restore frozen-source into a new research workspace with the external official baseline and unchanged benchmark. Pinned stock/Mode50 images and private metered Luna credentials are external. Historical Mode50 image identity is not established. The exact peer source and previous neutral runtime reproduction are included in the nested reference archive. Both populations have prior exposure and are separately reported. No current experimental outcome or raw benchmark/task/provider data is in this archive. All readiness evidence is neutral scripted execution or synthetic analysis. No paid inference has launched when this source archive is prepared.