a-stock-data

a-stock-data is a data toolkit designed to help AI coding assistants (like Claude Code or Codex) answer questions about Chinese A-share stocks — the stocks listed on mainland China's Shanghai and Shenzhen exchanges. The challenge it solves is that raw data for Chinese stocks is scattered across more than a dozen different sources, each with its own authentication quirks, header requirements, and API formats. This toolkit bundles all of that into a single "Skill" file, which is a structured document containing instructions plus embedded Python code that an AI assistant can read and use directly.

Once set up, you can ask your AI assistant plain questions like "what is the PE ratio of stock 688017" or "which stocks are attracting the most northbound capital today" (northbound capital refers to money flowing from Hong Kong into mainland Chinese markets through cross-border trading programs), and the assistant knows how to fetch and return the data. The toolkit covers 28 data endpoints across 7 layers: real-time price and candlestick charts, analyst research reports, sector and theme signals, margin trading and block trade data, company news, financial statements, and regulatory announcements.

The project is a self-contained Python toolkit requiring no third-party data wrapper libraries — it connects directly to the underlying HTTP APIs of Chinese financial data providers. It is compatible with any AI coding assistant that supports context injection, not just one specific tool. The full README is longer than what was provided.