About this skill

Problem

A-share traders often face fragmented data pipelines: real-time quotes, announcements, news, financial statements, fund flows, and macro indicators live in different sources. Tencent endpoints provide fast snapshots but limited depth. Some AkShare Eastmoney endpoints can be rejected, while Wind covers deeper financial, technical, index, and macro data but requires careful usage.

How the skill works

The skill organizes data sources by task and uses data_router.py for intelligent routing:
- Real-time quotes: prefer Tencent endpoints for A-share indices, watchlists, US indices, commodities, and crypto snapshots.
- Announcements: use ftshare-announcement-data for date-based lists, with Wind RAG as a semantic-search fallback.
- Deep data: use Wind servers such as stock_data, index_data, fund_data, bond_data, financial_docs, and economic_data for financials, technical indicators, sector data, macro series, and RAG queries.
- Fallback coverage: route news, sector, or fund-flow requests to WebSearch when primary sources are unavailable.
The router scores sources by availability, timeliness, and quality, then degrades gracefully on quota, timeout, or parameter errors.

Boundaries and notes

This is suitable for trading scripts, scheduled jobs, and AI-agent integrations that need stable market data. Wind output must be attributed, batch queries should call one instrument at a time, and CLI commands need to run from the skill directory. Free endpoints should not be treated as the sole source.

Screenshots

Use Cases

Build A-share watchlist snapshots by pulling index and stock quotes from Tencent endpoints with retry handling.
Run scheduled jobs for stock announcements, preferring ftshare and falling back to Wind announcement RAG on failure.
Expose financial statements, technical indicators, and macro series to an AI agent using scored source routing.
Monitor sector PE percentiles and ETF discounts/premiums by querying Wind index_data or fund_data in trading scripts.

Best For

Engineers maintaining A-share trading scripts who need one routing layer for quotes, announcements, and financials.
AI-agent engineers building finance assistants who need stable fundamentals, macro series, and announcement RAG calls.
Quant data engineers comparing free endpoints with Wind coverage, latency, and field quality.
Research or risk ops teams running scheduled jobs for announcements, sector data, and fund-flow monitoring with fallback.