skill-news-intelligence · Industry Intelligence Briefing
skill-news-intelligence aggregates in-depth coverage from Chinese industry media: it pulls deep articles and flash updates from 11 sources — including Huxiu, 36Kr, and woshipm — filters them by the creator's track, and produces a structured Chinese briefing (headlines, one-line summaries, and insight) plus 3–5 actionable topic angles.
Five preset daily briefings are built in (creator, tech, business, AI, trending topics). Sources that fail to fetch are skipped without breaking the briefing.
Example invocation: "Give me today's AI industry briefing."
Full brief
Positioning
skill-news-intelligence is the creator's industry intelligence desk: it pulls depth coverage from Chinese industry media, filters by track, and produces a Chinese briefing with executable topic angles. It answers "what happened in the industry lately, and what's worth digging into" at daily depth. Minute-level trending lists and quick second-creation belong to skill-trending-topics — this skill never scrapes trending charts.
Core capabilities
- 11 sources: woshipm, Huxiu, TMTPost, GeekPark, SSPAI, InfoQ China, AIHOT, 36Kr, WallStreetCN, Tencent News, ThePaper; fetchable in any combination by source key.
- Track filtering: self-media/ops, tech & gadgets, business & finance, AI, current topics; with a profile, the track is extracted automatically and topic suggestions match the creator's platforms.
- Depth modes: quick scan (titles + summaries) or deep analysis (downloads full text); quick is the default.
- Briefing format: headline scan (linked title, source, timestamp, one-line summary, background/impact/creator-relevance insight) plus 3–5 actionable topics based on the day's coverage (angle + suggested platform/format).
- Preset dailies:
creator,tech_digital,business,ai,topics. - Anti-hallucination: only fetched data is used, nothing invented; missing timestamps marked "unknown"; low-relevance filler is flagged.
- Fail-open: a failing source returns empty and is skipped — the briefing still ships.
Workflow
- Pick sources: by track, or
--list-sourcesto see all - Fetch:
fetch_news.py(--source,--limit,--keyword,--deep) - Compose: Chinese briefing per the rules (track filtering, topic angles, filler flags)
- Save:
outputs/资讯简报/news-briefing-{date}.md
Inputs & outputs
| Input | Notes |
|---|---|
| Track / field | Optional — self-media, tech, business, AI, topics (auto from profile) |
| Keyword | Optional — e.g. "AI", "funding" (title filtering) |
| Depth | Optional — quick scan / deep analysis (default: quick) |
| Volume | Optional — 5–15 items per source (default: 10) |
Output: Chinese intelligence briefing in Markdown (headline scan + topic angles).
Boundaries with adjacent skills
- skill-trending-topics: real-time trending lists from Weibo/Douyin/Zhihu/Toutiao/Bilibili at minute-level freshness, for trend-chasing and quick second-creation.
- skill-algorithm-updates: tracks only platform algorithm and policy mechanics.
Fit
- Creators building a daily reserve of depth topics and industry observations
- Track-watchers following key events in their field
- Media operators assembling editorial meeting material
- Accounts publishing regular industry commentary
Before you start
- Requires Python 3.8+,
requests,beautifulsoup4. - Supplying a track or keyword materially improves relevance; without a profile, general mode or an explicit track works.
- Deep mode downloads full text and takes longer; some sources (36Kr, WallStreetCN, Tencent) use dedicated fetchers, not generic RSS.
skill-news-intelligence is part of the Aiglade Skill library. Invoke it from the Aiglade chat box in plain language.