Quantitative Trading Engine Starter Edition

What problem it solves

This skill is for engineers who want screening, monitoring, and analysis in one workflow: it combines Eastmoney data, mx-xuangu filtering, and a local FastAPI decision service, reducing manual hand-offs between scripts. The default output is signals, scores, and recommendations rather than automated human-like decision-making.

How it works

The system has four roles: the data layer pulls market information, the monitoring layer runs conditional screening and maintains a watchlist, the local model layer analyzes candidates that pass evaluation, and the output layer returns recommendations or optional simulated trading. A typical flow is to convert a natural-language condition into screened results via /api/v1/screen, score the watchlist via /api/v1/evaluate, and send ready_for_model candidates to /api/v1/analyze, /api/v1/scan, or /api/v1/predict. State can be inspected through /api/v1/state and /api/v1/decision-log.

Boundaries

The skill assumes you provide and configure MX_APIKEY; it does not include free data authorization and does not guarantee third-party data accuracy. The starter edition is a basic decision engine for technical research and learning. Outputs are not investment advice, and compliance and risk remain the user's responsibility.