Methodology · v2.11

RESResiliencia

Composite organizational resilience score for {legal_name}. EXTRACTION RULES: return type="composite_signal" with one entry per dimension, each 0-100 + source_url. ANTI-PLACEHOLDER STRICT (Phase A_v2.5 audit): NEVER emit value=0 as placeholder for any sub-signal. If you cannot find evidence for a sub-signal, OMIT IT ENTIRELY from the signals array — do not include it with value=0. The server bails on all-zero composites as `composite_signal_all_zero_placeholder`. A composite with 2-3 real sub-signals is BETTER than 4 sub-signals where some are zero placeholders. Required sub-dimensions: (a) credit_rating_quality — 0-100 mapped from the average of Moody's + S&P + Fitch long-term issuer ratings (AAA/Aaa → 100, AA+/Aa1 → 95, AA → 88, A+ → 80, A → 72, BBB+ → 60, BBB → 50, below investment grade → 25). RECENCY HARD REQUIREMENT: the rating action MUST be from the LAST 24 MONTHS. A rating from 3+ years ago (e.g. a 2012 Fitch action cited in 2026) is STALE and FORBIDDEN — find the most recent affirmation/update/review on the agency site. If only a stale (>24mo) rating exists, OMIT credit_rating_quality (do NOT cite the old one). REQUIRED SOURCE: the source_url MUST be hosted on a rating-agency domain — `moodys.com`, `spglobal.com` / `standardandpoors.com`, or `fitchratings.com` — AND reference an action dated within 24 months. Citations to the company's own annual report PDF, IR page, or third-party summary are FORBIDDEN for this sub-signal. If you cannot find a recent rating-agency URL, OMIT credit_rating_quality from the signals array (do NOT emit value=0, do NOT cite a stale rating). WORKED EXAMPLE PATTERN: an issuer with Moody's Aa1 + S&P AA+ + Fitch AA+ maps to ~93, sourced from moodys.com/credit-ratings/<issuer>, spglobal.com/ratings/en/research-insights, fitchratings.com/entity/<issuer>. (b) outlook_stability — 0-100 from rating outlook consensus (stable → 90, positive → 95, negative → 50, watch-negative → 35); same source-host requirement as credit_rating_quality. (c) liquidity_buffer — 0-100 mapped from current ratio / quick ratio (>2 → 95, 1.5-2 → 85, 1-1.5 → 65, <1 → 40); source can be the company's most recent annual filing balance sheet OR an independent ratios page (stockanalysis.com / macrotrends.net / gurufocus.com). NOTE for BANKS: the traditional current-ratio is not meaningful for banks — use CET1 ratio (>15% → 95, 12-15% → 80, 10-12% → 65, <10% → 35) sourced from the bank's Pillar 3 disclosure or annual report. (d) geographic_diversification — 0-100 from share of revenue from outside single largest geography (>50% diversified → 90, 30-50% → 70, <30% → 45); source is the annual filing segment information note OR an independent segment-breakdown page. WORKED EXAMPLE PATTERN: ~55% non-domestic share → 85. Each signals[].value MUST be a real assessment based on cited evidence with the required source host. DO NOT emit a single scalar like "3 ratings" — the previous spec mapped a count to 30 which is structurally wrong for a top-credit-rated mega-cap. If NO sub-signal can be sourced (no recent rating, no public liquidity / segment data — typical for PMI / non-rated private companies), return type="absent" reason="not_disclosed" — NEVER fabricate a resilience score from weak evidence. Server applies composite_signal_score.