cclank/news-aggregator-skill

Overview

This skill is a comprehensive news aggregator that fetches, filters, and deeply analyzes real-time content from eight major sources: Hacker News, GitHub Trending, Product Hunt, 36Kr, Tencent News, WallStreetCN, V2EX, and Weibo. It supports broad global scans, targeted source queries, and deep fetching that downloads full article text for richer interpretation. The skill expands simple keywords automatically to cover related terms and returns structured JSON with optional deep content fields. It is optimized for daily scans, tech briefings, finance updates, and interpretive reports of hot topics.

How this skill works

The skill fetches top items from selected sources using flexible arguments: source, limit, keyword, and a --deep flag to extract article body. For general trend hunting, use the Global Scan mode which fetches broadly and relies on semantic filtering to surface high-value items. Keywords provided by users are automatically expanded into domain-aware lists to improve recall across sources. Output is a JSON array where deep mode includes a content field for each article and metadata for downstream analysis.

When to use it

• Daily briefings that cover tech, finance, and social trends across multiple Chinese and global sources.
• Targeted investigations when you need deep reading & interpretation of specific hot topics.
• Quick scans for top GitHub Trending repositories or Product Hunt launches.
• Time-windowed reports (past X hours) where filling sparse results with high-value items is acceptable.
• Research workflows that require article text for semantic summarization or downstream AI analysis.

Best practices

• Use --deep when you need article bodies for thorough interpretation or citation.
• Provide short topic keywords; the skill will expand them automatically—use specific terms only for unique queries.
• For comprehensive trend discovery, run Global Scan with higher limits and then apply semantic filtering.
• When requesting a time window, expect smart-fill behavior to ensure at least five meaningful items.
• For GitHub Trending, treat results as authoritative lists—do not expect smart-fill or time-window widening.

Example use cases

• Morning tech digest: run a Global Scan with --limit 15 --deep and generate a concise inbox-ready briefing.
• AI research: request keywords like 'AI' and receive expanded terms (LLM,GPT,RAG) plus deep analysis of each article.
• Finance update: fetch WallStreetCN and Tencent News items for the past 6 hours and get prioritized high-heat events.
• Product scouting: pull Product Hunt and GitHub Trending to spot promising launches and repositories.
• Incident deep-dive: fetch deep content for a breaking story to produce a 2–3 bullet analytical interpretation per item.