What problem does it solve? Financial research often relies on a single data source, which risks propagating errors from GAAP vs Non-GAAP differences, currency conversion mismatches, or stale platform data. This Skill enforces a rule that every key financial figure must come from two independent sources, with discrepancies above 1% explicitly flagged. ## Core Features & Use Cases - Dual-Source Cross-Validation: Defines primary and secondary data sources per market (macrotrends/stockanalysis for US stocks, aastocks for HK stocks, eastmoney/cninfo for A-shares, FinMind/Goodinfo for Taiwan stocks) and computes error rates with tiered handling (≤1% pass, 1-5% warn, >5% require original filing verification). - Taiwan Stock Data Toolkit: Ships a zero-dependency Python script (twstock_data.py) wrapping the FinMind API for quotes, valuation, 5-year financials, monthly revenue, and dividends, with built-in market-cap verification. - Price Adjustment Rules: Standardizes the use of forward-adjusted prices for historical analysis and backward-adjusted prices for total return calculations to prevent distorted historical comparisons. - Use Case: When researching TSMC, run the twstock_data.py script to pull financials from FinMind, cross-check against Goodinfo, and present revenue figures with both sources and the computed error rate annotated. ## Quick Start Ask the agent to research a company's financials, for example: "Get TSMC's last 5 years of revenue and net income, cross-validate the data from two independent sources, and flag any discrepancies over 1%."
