Add product_scoring.py: 10 deterministic advisory product scores ∈[0,1]
(market_fit/demand/supply_risk/future_competition/price_feasibility/infra_fit/
mortgage_sensitivity/differentiation/commercial/confidence), each derived from
already-built forecasting + Site Finder services, graceful-None on thin data,
§16-style RU reason per score. supply_risk/future_competition/mortgage_sensitivity
inverted (high-bad→low). Weighted overall renormalized over available scores
(§9.5 dispatch); commercial/infra degrade to None (not 0). Advisory cap inherited.
Pure scorers unit-tested without DB; orchestrator via @patch of all backers. 101 tests.
