Every Class H or Class M issue gets a Divisiveness Index (DI) from 0 to 100, built from how evenly opinion splits (polarization), how many distinct participants engaged (volume), how long the issue stayed contested (persistence), and how much disagreement exists across rater or endorser clusters (cluster_split). An issue lists once DI clears a threshold (currently 60.0); below it, the issue simply waits and can be rescored as new engagement arrives.
Stage Q asks a validation pool whether a scored issue is divisive enough to list: QUALIFY, REJECT, or DEFER (requeued once with an enlarged pool, then REJECT). Stage T opens a market and a separate truth pool once listed, settling a TRUE, FALSE, or UNRESOLVED verdict. Both stages use the same staked, commit-reveal, Trial-then-Full mechanism as the rest of Open Standing: a no-slashing Trial round earns entry to a pro-rata-slashing Full round.
Class H exists to test the mechanism, not to adjudicate people. No claim naming or imputing conduct to an identifiable natural person or organization is ever scored, pooled, served, or stored in a served field. The gate is automatic and fail-closed: anything it cannot classify is rejected, and a rejected submission serves nothing at all.