The Market Scanner Toolkit is a full-market scanner, a sector-breadth regime gauge, and a study harness for killing bad trade ideas cheaply — with the engineering guide that explains why each number is there and how it lies when read alone.
Strongest 20-day momentum that is NOT already overextended — the extension cap keeps yesterday's blow-off out of tomorrow's buy list.
Names above their 200-day having a bad week (RSI ≤ 35) — pullbacks in uptrends, not falling knives; the trend filter is what separates them.
Extended past the cap or RSI ≥ 75 — the risk-off list to check against what you own.
% of the 11 sector ETFs above their own 50-day. The index can be dragged up by three giants while eight sectors roll over — breadth sees it, and the verdict tells you how to read the other three tables.
All vendor contact in one file (FMP implementation included): call-budget throttle, plan-gate detection, the endpoint-family lesson that cost us an evening.
Test calendar and rotation hunches over years of data in twenty lines — no lookahead, win rates beside averages, and a printed reminder about what a positive result does and doesn't mean.
scan.pythe scanner: metrics → LEADERS / OVERSOLD / STRETCHED + regime, as Markdown + JSON reportsprovider.pythe vendor seam (FMP impl): throttle, plan gates, swap-one-file portabilityindicators.pySMA · EMA · RSI · ATR · extension — stdlib, shared with our engine Blueprintseasonality.pythe study harness: monthly seasonality + the failed sector-chase, method rules enforceddata/config.jsonyour curated universe + extension capcrontab.examplescan after each close; studies weeklyPython 3.10+, standard library only. Needs a market-data key (a ~$20–30/mo FMP plan covers everything the toolkit uses; the guide covers the traps of cheaper tiers).