This is a quantitative trading framework that connects Claude to FinLab's market data and backtesting engine across five markets (Taiwan, US, Korea, Japan, Hong Kong). It handles the full loop: pull price and fundamental data with `data.get()`, build factor conditions using DataFrame methods like `.rise()` and `.is_largest()`, combine them into position signals, then run `sim()` to backtest with realistic fees and slippage. The US market support is detailed, covering filing-date-aligned fundamentals and ETF rotation strategies. You'll need a FinLab API token (free tier gives 500MB daily). The workflow is clean if you're comfortable with pandas-style operations, though the documentation leans heavily on Taiwan market examples.
npx -y skills add koreal6803/finlab-ai --skill finlab --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills