# Financial Research Agent Eval: Korea Cheap Screen Deletion

I built a small public-market research eval for LLM/agent builders.

Task: given Korean cheap-screen candidates, can the agent delete weak ideas using KRX/KIND, DART, liquidity/access, minority economics, governance, and theme-to-P&L checks before inventing a thesis?

Core survivor stack: Poongsan, BHI, Woojin.
Moonshot drawer: Innospace, Mobiis, CellBion, IntoCell, SBB Tech.

Agent assets: https://barneywohl.github.io/korea-research-map-agents/
Repo: https://github.com/barneywohl/korea-research-map-agents

Not investment advice. This is an eval/design prompt for financial agents.