What problem does it solve? It turns open-ended investment idea or theme requests into a structured, reproducible screening process, producing candidate universes, scorecards, and shortlists instead of ad-hoc stock tips. ## Core Features & Use Cases - Criteria Structuring: Parses the user's raw screening prompt into explicit inclusion/exclusion conditions, market scope, and reference date saved to screen-criteria.md. - Candidate Universe & Scoring: Builds a candidate list from DART/KRX, yfinance, IR, and exchange sources, then scores each on six dimensions (Thesis Fit, Market Tailwind, Financial Quality, Valuation/Risk-Reward, Catalyst Clarity, Data Confidence) with a 0-5 scale. - Shortlist with Preliminary Ratings: Produces a 3-10 name shortlist with preliminary Buy-to-Sell ratings, key risks, and recommended next analysis steps (/analyze, /comps, /dcf, /earnings). - Use Case: A user asks for Korean semiconductor suppliers benefiting from AI demand; the skill records the criteria, gathers candidates, scores them, and writes a ranked shortlist to ${ACTIVE_WORKSPACE}/00_screen/ for follow-up deep-dive reports. ## Quick Start Ask the agent to screen investment ideas for a theme such as US grid-infrastructure beneficiaries and save the scored shortlist to the active workspace.
