Records Sandbox: correction, regression and recovery Recorded 2026-11-20 on a local build with synthetic data and AI-cloned English narration. 00:00:00.000 · Starting check This is the Records Sandbox. We start with a synthetic oxygen saturation observation. The draft declares both the oxygen saturation and laboratory profiles. The laboratory category is present. The check flags the missing vital signs category. 00:00:22.570 · Reviewed correction The reviewed change adds vital signs and preserves laboratory. We run the check again and inspect the resolved findings. The missing vital signs requirement is now resolved. The measurement stays at eighty two percent. 00:00:42.540 · Intentional regression To test for a regression, remove the laboratory category from the temporary draft and check the same profiles again. The vital signs requirement is still satisfied. The check now reports a new finding for the missing laboratory category. 00:01:03.210 · Recovery Restore both categories and check the draft again. Neither category requirement is missing. The first check remains the comparison baseline in this Sandbox session. Find the measurement and the current validation result side by side. 00:01:27.680 · When a check is incomplete A separate follow-up adds an extension with no resolvable definition. Records reports missing validation basis and withholds the comparison. These checks use a temporary synthetic draft. The source server is untouched.