Financial Report Downloader

Problem: locating report PDFs is brittle

Listing company announcement pages often load dynamically, making raw HTML scraping for announcementId or final PDF links fragile. A-share, HK, and Beijing Stock Exchange sources differ in orgId formats, search parameters, and download domains. Manually distinguishing official annual reports from notices, summaries, revised versions, or English versions is slow and error-prone.

How it works

The skill condenses the workflow into three routes:
- Shanghai/Shenzhen/HK: query the Cninfo announcement API using parameters such as stock, searchkey, and category, then extract announcementId. If the response includes adjunctUrl, it can be used to build the direct PDF link; otherwise agent-browser renders the detail page and extracts the static.cninfo.com.cn link from the anchor tag.
- Beijing Stock Exchange: because Cninfo does not cover this exchange, it switches to the Eastmoney API and downloads PDFs directly from pdf.dfcfw.com.
- Output normalization: files are named like {year}年报.pdf or {year}Q1.pdf, and default output is grouped under cninfo_reports/ for batch organization.

The workflow accounts for common edge cases: the Shenzhen prefix is gssz, not gszs; some orgId values are numeric or prefixed with GD or nssc; searches for “annual report” may return disclosure notices, summaries, revisions, or English versions; the publication year can lag the reporting period; and Cninfo page loading should avoid networkidle hangs.

Boundaries

It is suited for bulk retrieval of already-published annual and quarterly report PDFs for A-share and HK-listed companies. It does not perform financial statement parsing, indicator extraction, or compliance review. If APIs are rate-limited, anti-bot protected, or PDF downloads fail, retries, proxies, or manual title verification may still be required.