Research Report Query Assistant

Problem and Use Cases

When you need quick research around industries, individual stocks, strategy reports, macro notes, and broker morning briefs, manually checking report lists, judging page counts, extracting key points, and downloading PDFs becomes tedious. The Research Report Query Assistant targets this workflow and completes query and download steps over HTTP, without requiring an extra CLI. It is useful for engineering-side research collection, industry comparison, individual-stock rating screening, and saving report summaries and PDF files to a local directory.

How It Works

The skill splits request paths by report type: industry, strategy, macro, and morning brief requests usually use GET calls to return JSON lists; individual-stock reports use bash curl to send a POST request because stock parameters are required. List responses include fields such as title, orgSName, publishDate, emRatingName, encodeUrl, and attachPages, which support title filtering, rating filtering, and page-count checks. For summaries, the skill builds a detail-page URL from encodeUrl, extracts the report abstract and PDF download link, then uses curl to save the file.

Boundaries and Notes

The skill depends on a clear report type, date range, and either an industry code or a six-digit stock code. Strategy, macro, and morning brief lists may not return page counts directly, so the skill may need to inspect detail-page length. It does not support HK-listed stocks, and it applies constraints to API intervals, PDF download frequency, and downloads per session, so rate-limit guidance should be respected. The data is intended for learning and research use.