CCTV News Broadcast Daily Collector

Problem It Solves

Manual daily summaries of CCTV's News Broadcast are often blocked by fragmented sources: the same day's script may appear on CCTV, Cailianshe, Xinhua, or People's Daily pages, while main items and short bulletins need to be separated. Adding public reaction makes the job heavier: comments may be scattered across Weibo, Toutiao, Zhihu, stock forums, and news comment sections, and the writer must identify the source before summarizing it. This skill turns that workflow into a repeatable task: collect the day's broadcast items, search public commentary, and output a structured Markdown report.

How It Works

The workflow is organized around three stages:
- Broadcast collection: uses WebSearch to locate the day's text script, prioritizing reliable pages such as ysxw.cctv.cn, cls.cn, and tv.cctv.cn, then extracts the main items in broadcast order.
- Comment retrieval: for items with higher public-interest likelihood, runs parallel searches using terms like netizen comments and public discussion, then identifies the source platform, such as Weibo, Toutiao, Zhihu, or financial community threads.
- Report generation: writes a dated Markdown file using references/output-template.md, typically under /新闻联播日报/YYYY-MM-DD/, with each summary kept concise and the file ending with source references and generation time.

Boundaries

The skill is best used for daily news summarization and public-sentiment direction tracking, not formal editorial production or deep opinion research. If the broadcast has not aired, the script cannot be found, or a topic has no public discussion, the output should say so plainly instead of inventing content. Comments must be grounded in identifiable sources, and the skill should not present the author's judgment as public reaction.