Financial Report Download and Collection

Problem

When a listed company's financial report needs to be collected by company name or ticker, and then archived into an ima knowledge base, manual searching, filtering, PDF download, folder placement, and duplicate checking can be repetitive. This skill turns that process into a fixed workflow: query Juchao Info announcements, confirm the target file, then complete download and archive serially.

How It Works

The skill accepts a company name or ticker, report type, and target knowledge base or folder. Key steps include:
- Build queries using Annual Report, Q1 Report, Semi-annual Report, or mapped equivalents
- Parse announcements[], filter for adjunctType=PDF and matching titles
- Present candidate reports with title, date, and size when multiple matches exist
- Download and validate the PDF before entering the ima upload phase
- Check credentials, resolve the knowledge base and folder, append a timestamp on name conflict, then create media, upload via COS, and add knowledge through ima_api

Boundaries

The workflow is strictly serial and stops on failure. Upload depends on ima-skill and local credentials in ~/.config/ima/client_id and ~/.config/ima/api_key. If the target folder does not exist, the report falls back to the knowledge base root; if the company name, ticker, or date range is inaccurate, results may be missing.