METH-02
Coverage
Published per tool only when the API returns source and coverage metadata.
No platform-wide accuracy extrapolation.Method ledger · versioned truth
A result is useful only with its method and limits.
Method records belong beside the tool they govern. Missing coverage, stale sources and withdrawn methods do not become optimistic copy.
Current public posture
Six questions every method must answer.
No cohort, product or result is generalized across AspireACE.
METH-03
Known gaps
Empty, stale and unavailable source states fail closed.
No synthetic result fills a missing official source.METH-04
Disagreement
Question feedback follows a signed attempt; predictor output remains a planning signal.
A verified human escalation owner is still required.METH-05
Language
Current public serving is English-first.
No multilingual performance claim is published.METH-06
Accessibility
Native controls, visible focus and reduced-motion rules are part of the website layer.
Assistive-technology production audits are not yet claimed.METH-07
Fairness
No fairness score or cohort result is published without study design and review.
Research outcome claims remain gated.Research gate
Outcome claims remain withheld.
Cohort size, study design, confidence, limitations, reviewer, source and conflict disclosure are all required before publication.