kdaip/echobot/news

News

Fetch and present the latest headlines using the fetch_web_page tool.

Categories and Sources

Workflow

1. Identify what the user wants

• Single category: fetch that category's URL.
• Multiple categories: fetch all relevant URLs in parallel (call fetch_web_page multiple times in the same response).
• No category specified: default to General (新华网) or ask if ambiguous.

2. Fetch the page

Call fetch_web_page with a generous max_chars to capture enough headlines:

{
  "url": "https://36kr.com/",
  "max_chars": 10000
}

Use max_chars between 8000–12000 for news homepages — smaller values cut off too many headlines.

3. Extract headlines

From the returned text, look for repeating patterns: headline titles, brief descriptions, timestamps. Most Chinese news sites return plain-text content with clear headline structures after HTML is stripped.

Focus on:

• The headline (title of the story)
• A one or two sentence summary (from the teaser or first paragraph)
• The date/time if visible in the text
• The source name

4. Present results

Format your reply like this (adapt to Chinese or English based on the user's language):

📰 [Category] 最新资讯 — 来源：[Source Name]

1. [Headline]

[One or two sentence summary.] (date/time if available)

1. [Headline]

[Summary.]

(more headlines...)

🔗 查看完整报道：[Source URL]

• List 5–10 headlines per category. More is fine if the user asked for an overview.
• If the user asked for multiple categories, present each as a separate section.
• Match the user's language: reply in Chinese if they asked in Chinese, English if in English.

Error handling

Tips

• Fetch multiple categories in parallel (same response, multiple tool calls) to respond faster.
• Don't repeat the same headline across categories if sources overlap.
• Always include the source URL in your reply so the user can read the full story.
• If the user specifies a city or region (e.g., "北京新闻", "上海财经"), note this in your reply but use the same national sources — regional editions are not in this list.