Daily Potential Web Novel Recommendations SBTI Edition

Problem

Qidian Sanjiang rankings update often, but the book list is scattered across web pages. Readers tracking new releases usually repeat the same cleanup work: identifying promising debut works, checking genre fit, avoiding repeats, and summarizing why a book is worth reading. Classic web novels add another layer: 10k/100k-tier status, IP adaptations, and overseas popularity need separate verification. This skill turns that workflow into a repeatable recommendation process.

How It Works

• Data source and caching: It prefers public Qidian Graph data. First request fetches the full list; later requests read local cache. The docs say the same day's data is fetched once to reduce source-site pressure.
• New-book recommendations: It selects titles from the Sanjiang list and outputs title, author, category, summary, and Qidian link. The report must use the returned qidian_url value so the _trace attribution parameter is not lost.
• SBTI matching: It can filter by code, Chinese name, or free-form personality description, such as “lying flat”, “introvert”, or “overthinking”, using exact, keyword, and fuzzy matching.
• Classic mode: It covers 10k and 100k-tier works, highlights tier labels, IP derivatives, and overseas exposure, and can use built-in data plus incremental checks to reduce wait time.
• Deduplication: It excludes the previous recommendation using local history and falls back to the full candidate pool if needed, so daily picks do not repeat.

Boundaries and Notes

• Data comes from a third-party source and may lag official Qidian by 1-2 days; missing dates fall back to the last 7 days.
• Results depend on daily rankings, cache state, and personality labels, so it is not suitable for real-time official data or complete catalog coverage.
• It fits daily new-book discovery, personality-based selection, and curated classic web novel reports, but should not be treated as the sole source for exact copyright, sales, or publication data.