proof-of-work-scoringlisted

# Proof-of-Work Scoring ## Overview Score candidates against a Role DNA using **evidence-gated, weighted, auditable** scoring. Hard dealbreakers gate first; surviving candidates get per-dimension judgments that combine into a reproducible composite via `csi_composite_score`. Feedback (accept/reject) calibrates the weights over time via `csi_calibrate_weights`. Core principle: **every dimension score must cite evidence**, and the arithmetic is deterministic so two runs on the same inputs produce the same number. ## When to Use - Ranking a sourced pipeline against a role - Rating a single candidate and explaining why - Tuning weights after you've recorded real advance/reject decisions **Requires:** a Role DNA (`role-dna`) for weights, must-haves, and dealbreakers; and candidate evidence (`evidence-sourcing` / `candidate-dossier`). Don't score on a bare name. ## Configuration If no Role DNA is supplied, fall back to `.calibratr/config.json` (working directory) for `scoring.default_weights`, `scoring.threshold`, `scoring.dealbreakers`, and `company.values` / `target_signals` / `exclusion_signals` as the Company DNA for `dna_fit`. Absent → use the built-in defaults (weights `{role_fit:55, dna_fit:30, impact:15}`, threshold 60). See CONFIGURATION.md. ## Workflow 1. **Dealbreaker gate FIRST.** Check the candidate against Role DNA `dealbreakers`. If any `severity: "absolute"` anti-pattern is clearly met by the evidence → call `csi_composite_score` with `dealbreaker_tr