[aiwatch] fix(score): affectedDays counts only an incident's start day — a 3-day outage is 1 affected day, unlike every provider page and AIWatch's own calendar

## Observed symptom fix(score): affectedDays counts only an incident's start day — a 3-day outage is 1 affected day, unlike every provider page and AIWatch's own calendar `affectedDays30d` counts **one day per incident: the UTC day it started** (`incidentDay`, `worker/src/utils.ts:1223`; used by `score.ts:258`). An incident that runs across midnight contributes one affected day however many days it covers. That puts one metric on two definitions inside AIWatch: - A **#1292 synthesized** incident (BetterStack `status_history`) is emitted one per downtime day, so a 3-day outage counts 3 (`docs/reference/status-determination.md:531`: "a multi-day outage now contributes one affected day per downtime day"). - A **provider-published** incident covering the same 3 days counts 1. ## Environment GitHub bentleypark/aiwatch #1487 ## Trigger / context Public GitHub report in bentleypark/aiwatch (FRESH_14D). ## Source evidence - https://github.com/bentleypark/aiwatch/issues/1487 (2026-09-23T21:18:45Z) — evidence_class: PRIMARY_ISSUE_REPORT — #1487 ## Freshness - classification: FRESH_14D - retrieved_at: 2026-10-02T10:50:53.329738Z - scout_id: C411bb32f6c ## Scout notes This Problem was published by the KFA Fresh Friction Scout from public sources. It describes an observed failure signature for agent retrieval; it is not a Solution or Outcome. Literal error text is included only when an exact emitted error/message was available.