AI Content Collector

Problem

Collecting AI and automotive industry updates often returns forums, social posts, and product pages mixed together. If a table is built only from search snippets, dates, sources, and claims are hard to audit. AI Content Collector turns source filtering, article fetching, cross-checking, and structured output into a repeatable workflow.

Core workflow

The skill first parses the request into a DATE_START to DATE_END range, then prioritizes authoritative outlets such as IT Home, The Paper, 36Kr, and Jiqizhixin using site: searches. When Chinese results are thin, it falls back to fetching news-listing pages instead of simply switching to English media. Every candidate item must be opened from the original article to extract the date, actor, key metric, and primary link; items from Zhihu, Weibo, forums, or Toutiao are skipped. The output includes category, source or issuer, title, publish date, summary, primary link, and supporting links.

Boundaries

It is suited to weekly briefs, competitor updates, and industry research drafts, not real-time monitoring, legal due diligence, or high-stakes sentiment analysis. English sources remain supplementary; major events need cross-validation from distinct domains. Records with vague dates, single unverifiable sources, or inaccessible original pages are dropped.