fix: verify and rebalance question data

feat: remove Supabase, bundle questions locally, use JSON DB + API

fix: cancel button actually cancels matchmaking

fix: batch of 5 stability fixes

fix(spark): Batches 2 + 3 — atomic ELO, real DB IDs, polish

fix(spark): Batch 1 — practice ELO leak, schema alignment, RLS tightening

feat(spark): add Challenge a Friend + shareable verdict card

feat(spark): add custom match invite link with waiting screen and copy button

feat(spark): rewrite landing page with hero preview, CTAs, leaderboard, deferred spectrum

feat(spark): add landing page queries for preview, leaderboard, recent debates

docs: Spark landing page redesign implementation plan

docs: Spark landing page redesign spec

feat(spark): open matchmaking + persist practice debates to DB

feat(spark): practice mode awards half ELO

fix: increase Node memory limit for HF build

feat(spark): complete debate arena rework to conversational flow

fix(spark): update downstream components for new RoundData field names

feat(spark): rewrite useDebate for conversational flow with commentary + challenge

feat(spark): add challenge persistence and update round schema

feat(spark): update judge with arc-based criteria and challenge support

feat(spark): add inter-round commentary API endpoint

feat(spark): swap AI arguments to Claude Sonnet 4.6 with 150-word cap

feat(spark): rewrite state machine for conversational flow

feat(spark): update types for conversational debate flow

docs: Spark debate rework implementation plan

docs: Spark debate arena rework design spec

fix: leaderboard tracking, score bugs, mobile admin, OG image

feat: make multiplayer matchmaking snappy and engaging

feat: improve UX and player tracking

fix: simplify topic selection — host picks, guest receives

fix: resolve multiplayer matchmaking race condition in Spark debates

feat: redesign Spark landing page with arena aesthetic

feat: add OG social images for Triviaverse, Spark, and Game Night

fix: judge now scores on debate technique -- rebuttals, pressure, consistency

fix: judge evaluates delivery/reasoning/engagement, not crowd appeal

fix: simplify judge and stop AI opponent from fabricating anecdotes

fix: judge now penalizes repetitive emotional appeals and rewards specificity

fix: ensure "Different topics" generates 3 fully new topics

refactor: rewrite judge prompt for crowd-persuasion scoring

fix: completely rewrite fact-checker to only flag objectively false claims

fix: fact-checker must not cite single org as sole authority

fix: fact-checker bias and AI word limit

fix: AI opponent takes opposite side based on spectrum + judge anti-style bias

fix: judge must track cross-round references and contradictions

feat: add full debate transcript and expandable rounds to verdict

fix: add explicit bias prevention rules to judge prompt

fix: overhaul judge and AI opponent for substantive debate

feat: allow users to suggest custom debate topics

fix: make debate topics substantive and real-world

fix: resolve all type errors from spectrum redesign