daishin-report-search

To retrieve the latest Daishin Securities research reports as structured data, this Skill fetches HTML pages from a public GitHub Pages mirror and returns machine-readable JSON metadata including titles, headings, text excerpts, and original URLs for downstream reuse.

Yes, you can search Daishin Securities reports by matching query terms against titles, headings, and body text. The Skill parses raw HTML content from public URLs and returns matching report metadata in a structured JSON format ready for agent consumption.

To fetch a specific Daishin Securities report using its timestamp ID, provide an ID like YYYYMMDDHHMMSS. The Skill locates timestamped HTML files via GitHub API tree listing and retrieves the original text and an optional companion explain page.

A GitHub token is optional when fetching HTML pages from the GitHub Pages mirror. The Skill supports optional GitHub token authentication to bypass API rate limits, using a robust fallback to the exact-file contents API when direct retrieval is unavailable.

The JSON output for securities research reports includes titles, headings, body text excerpts for summaries, indicators for Rating/Target table presence, and direct URLs for both the original report and the companion explain page, enabling immediate downstream agent reuse.

HTML parsing may fail when direct retrieval from GitHub Pages is unavailable due to access restrictions or broken links. The Skill implements a robust fallback mechanism using the GitHub exact-file contents API to ensure raw HTML fetching and text extraction succeed.