Findings & scores

A finding is one verified observation from named sources — never speculation.

Anatomy of a finding

{ "section": "Domain registration", // the category "finding": "Domain registered 1999, registrar MarkMonitor.", "risk": "LOW", // LOW | MED | HIGH "compliance": "PASS", // PASS | REVIEW | FAIL "sources": ["WHOIS"], // what backs it "input": "grupobimbo.com", // exactly what was queried "origin": "exploration" // baseline | exploration }

Risk levels

LEVEL	MEANS
RISK: LOW	Consistent with a legitimate, established business.
RISK: MED	Worth a human look — an inconsistency or a gap.
RISK: HIGH	A material adverse fact from an authoritative source.

Compliance

PASS / REVIEW / FAIL is the check's outcome against expectations; risk is what it implies. A check can PASS and still carry MED risk (e.g. a very young domain that is validly registered).

The two scores

Deterministic and auditable. The scores are computed from the findings list — not model output. The same findings always produce the same scores, and the score is explainable: "75, because two HIGH findings and two MED." riskScore aggregates risk levels; complianceScore starts at 100 and is reduced by failed and review-flagged checks.

Confidence vs. scores

confidence is about how well the investigation covered the entity (sources answered, claims corroborated). The scores are about what was found. A thin record can yield high risk with low confidence — read both.