Communitygithub.com

blaze10011/market-flow-desk

Quant research desk: 23 Python engines + 18 Claude Code agent skills modeling capital rotation — sector flows, thematic value chains, technical lenses, SEC EDGAR filings radar, and book-level risk. Decision support, not signals.

¿Qué es market-flow-desk?

market-flow-desk is a Claude Code agent skill that quant research desk: 23 Python engines + 18 Claude Code agent skills modeling capital rotation — sector flows, thematic value chains, technical lenses, SEC EDGAR filings radar, and book-level risk. Decision support, not signals.

Compatible conClaude Code~Codex CLI~Cursor
npx skills add blaze10011/market-flow-desk

Preguntar en tu IA favorita

Abre un nuevo chat con esta habilidad de agente ya precargada.

Documentación

Money Flow Index — volume-weighted RSI

RSI only sees price; MFI weighs volume too. Comparing the two separates a conviction move from a thin one — exactly the SOFI case in the brief (RSI fine, but volume 0.1× average → no real buying).

Steps

  1. Run from project root: python3 scripts/mfi.py [TICKERS...] (no args = all 10 traded names)
  2. Interpret only what it prints (MFI, RSI, and the gap between them):
    • aligned (|gap| ≤ 8) → the move has volume conviction; trust it more.
    • MFI well BELOW RSI → price rose on thin volume; weak/suspect, easy to reverse — don't chase.
    • MFI well ABOVE RSI → heavy accumulation the price hasn't fully reflected; constructive.
    • 80 overbought / <20 oversold, same scale as RSI.

  3. Deliver a one-line read per name: is today's move backed by volume or not, and what that does to the lean's confidence.

Use it to grade the CONVICTION of a signal from the other tools, not as a standalone trigger. Decision support, NOT advice. Data ~15-min delayed.

Skills relacionados