About this skill

Problem: fragmented and unstable A-share data sources

A-share quotes, research, fund flow, dragon-tiger lists, and announcements are scattered across TDX, Tencent, Eastmoney, THS, Sina, and CNInfo. Interfaces differ in GBK encoding, field positions, expiration behavior, empty responses, and even silent wrong-symbol returns. The engineering problem is not fetching one field, but turning several data sources into a routable, fallback-ready, throttled call path.

How it works: endpoint routing, fallbacks, and throttling

The skill exposes 44 endpoints for K-lines, order books, tick data, real-time prices, PE, PB, market cap, research PDFs, consensus EPS, northbound flow, concept membership, minute fund flow, dragon-tiger seats, daily dragon-tiger boards, lockup expiry, industry comparison, board fund flow, margin trading, block trades, holder changes, dividends, news, announcements, limit-up pools, ETF options, investor Q&A, and hot lists. It prefers mootdx and Tencent endpoints because they are less IP-risky, and uses Eastmoney only for exclusive data through em_get(), which serializes requests, reuses the session, and adds delay. When a primary source fails, dragon_tiger_backup(), fund_flow_backup(), and announcements_backup() can degrade to exchange, Sina, or market-specific announcement sources. It also normalizes tickers, applies market prefix rules, validates mootdx servers, and reduces ambiguity for codes such as 000001 and 000016.

Boundaries

It is suitable for A-share data querying and batch scripting, not for order execution or compliance risk control. Eastmoney endpoints are subject to QPS, concurrency, and UA risk controls, so batch jobs should increase EM_MIN_INTERVAL. Overseas IPs may time out on mootdx. Shenzhen Connect minute flow is reference-only. iwencai semantic search requires an API key.

Screenshots

Use Cases

Pull K-lines, order books, and tick data for multiple A-share tickers without hitting IP rate limits.
Batch-download individual and industry research PDFs and extract ratings, three-year EPS, and consensus estimates.
Audit same-day dragon-tiger seats, market-wide net buy rankings, and future lockup-expiry pressure.
Build a morning note with northbound minute flow, concept membership, and board-level fund flow.

Best For

Quant engineers building A-share strategies who need K-lines, order books, limit-up pools, and option fields.
Equity researchers covering China A-shares who need report PDFs, consensus EPS, industry comparison, and IR replies.
Data platform engineers who need to wrap multi-source A-share endpoints into fallback and throttled services.
Risk or compliance analysts checking dragon-tiger seats, lockup expiry, block trades, and holder changes.