priority-weight

Rank design audit clusters by scoring five priority dimensions: severity, reach, persistence, business impact, and cognitive cost. This generates a comparable per-dimension vector to guide downstream decision-making.

Multi-dimension priority scoring evaluates a reconciled cluster across five distinct dimensions using a 0–10 scale with calibration anchors. It yields a robust vector for comparing clusters without applying final weighting.

Apply priority weights by passing the ReconciledVerdict and cluster context into the scoring skill. It outputs raw per-dimension scores from 0–10, leaving final meta-weight calculations to external systems.

You can use custom meta-weights by applying them externally to the raw per-dimension scores. The skill intentionally leaves final weighting out, providing a robust score vector for your downstream scoring logic.

Priority scoring requires L5 reconciliation outputs, specifically the cluster context, quotes, and the ReconciledVerdict. These inputs provide the necessary tension data to score the five priority dimensions accurately.

You should not use automated priority scoring when final meta-weight alignment is needed within the tool itself, or when an L5 cluster lacks a reconciled verdict and sufficient context for the five-dimension evaluation.