Which AI model + skill builds the best on-brand Swiirl deck?

Same brief, same brand layer, measured cost and latency, deterministic brand gates, and a human 10-dimension rubric you can review and correct live.

Brief Cadence (de-identified · content-hashed) Brand Swiirl tokens + rules Quality 10-dim rubric /100 Efficiency $ + seconds (separate) Correct export overrides → aggregate

✎ Review & correct scores ◫ Compare decks ✦ Studio (BYOK, not scored) Rubric GitHub

Leaderboards

Dual boards

Primary: best run per model (phase-1 skill preferred when unreviewed). Secondary: best run per skill. Quality is human-corrected; cost/time never silently buy rank.

Matrix

Model × skill cells

Each cell is the canonical run for that pair. Click through to review.