This is a statistical arbitrage screener that finds cointegrated stock pairs for market-neutral trading. It runs correlation analysis and Augmented Dickey-Fuller tests to identify pairs with mean-reverting spreads, then calculates z-scores to signal entry and exit points. The workflow is thorough: pulls two years of price data, filters by correlation thresholds (0.70+), validates cointegration (p < 0.05), and calculates half-life for mean-reversion speed. You'd use this if you're building pairs trades in specific sectors or want hedged positions that profit from relative moves rather than directional bets. The methodology is solid and well-documented, though you'll need clean data and patience since it's designed for systematic strategies, not quick scalps.
npx skills add https://github.com/tradermonty/claude-trading-skills --skill pair-trade-screener