hot-topic-curationlisted

# Hot Topic Curation Use this skill when the task sounds like: - “做一个 AI 日报 / 热门话题专项” - “抓几个外部热点站聚合到站内” - “维护热点源、更新摘要页、顺手路由到对应专题” - “像 AIbase Daily 那样做一个长期维护的热点聚合页” This is a project-local editorial + ingestion skill. Prefer repo conventions over generic scraping habits. ## Core Outcome Produce or maintain a lightweight hot-topic aggregation flow that does all of the following: 1. fetch trusted public source pages or feeds 2. normalize them into a reusable local snapshot 3. keep the hub page updated 4. route worthwhile findings into the right site hotspot or topic pages 5. preserve source links and avoid full-text copying ## Workflow ### 1. Read the source policy first Before adding a new source, read: - `references/source-policy.md` - `references/route-hints.md` This skill is deliberately conservative: - prefer RSS / feed endpoints first - use HTML daily pages when the site exposes a stable public digest page - keep only title/link/date/one-line note level data in the local snapshot - never turn this into a full-content mirroring pipeline ### 2. Inspect the current source registry The canonical source registry is: - `site/data/hot-topic-sources.json` It defines which sources are watched, how they are parsed, and which site routes they usually feed. Do not hardcode a new source directly into prose if it should become a durable maintained source. Add it to the registry first. ### 3. Refresh the local snapshot Use the repo tool: ```bash python3 tools/fetch_h