Explainable by design
Methodology
A readiness/evidence score with a separate coverage measure—never a compliance probability.
Two numbers, one honest boundary
Readiness is earned weight divided by assessed weight. Coverage is assessed weight divided by the full 100 weighted points. If only one 15-point topic is assessed as evidence exists, readiness is 100% and coverage is 15%. Both facts matter.
readiness = earned assessed weight ÷ assessed weight
coverage = assessed weight ÷ 100
partial = 50% of category weight
unknown = excluded from assessed weight
Evidence streams do not collapse
Observed
A bounded public response signal, with detector caveats. None are produced while the fetch path is disabled.
Self-attested
A questionnaire answer. It can contribute to readiness but remains low-confidence until checked.
Document-derived
A browser-only parser result. It creates a review lead and never changes a score automatically.
Unknown
Missing or unassessed evidence. It reduces coverage but is not silently scored as failure.
Versions and change log
score-v1.1 preserves the original score-v1 arithmetic while making the assessed denominator explicit. review-topics-2.0 maps the canonical eight weighted categories plus two context questions. official-sources.2026-08-30 dates the legal-source snapshot.
- v1.1 — labels readiness as a percentage of assessed weight.
- v1.1 — separates public, self-attested, document-derived and unknown evidence.
- v1.1 — adds 30/60/90 evidence planning and export version fields.