The core tension to design around

The competition wants AI workflows that produce trustworthy knowledge bases grounded in real cases — making the provenance, structure, and assessment of knowledge transparent and traversable, across cases like COVID lab-leak origins, LHC/black-hole safety, and the egg/cholesterol nutrition literature. The sibling repo's own goal statement: expose "what is known, what is contested, what depends on what, and where the gaps are — without adjudicating truth."

That last clause points opposite to Commonplace-as-built. Commonplace is optimized to distill transferable, committed methodology for agents: title-as-claim, "do I still believe this?", the whole Popperian maintenance loop assumes the KB takes positions and defends them. Casework instead needs a mode that represents contestation faithfully and refuses to average it away. Mechanistic constraints make Popperian KB recommendations actionable warns why this matters: contradictions loaded into one context "aren't flagged; they're silently averaged." A casebook's value is keeping the disagreement structurally distinct instead of letting the model launder it into one confident answer.

So the highest-leverage additions are not gadgets but a register and a type surface for stance-neutral evidence maps, plus provenance strong enough to survive review. The idea documents linked from the workshop README are organized by the competition's three layers.