Author task-driven elicitation, collect responses against the toolbox kit, score via thepreference-modeler spoke. MVP: MaxDiff best-worst scaling.
preference-modeler
One per line. Use id | label or just a label (id is slugified).