hot_topics_selector

Original：🇨🇳 Chinese
Translated
2 scriptsChecked / no sensitive code detected

Financial Hot Topic Selection Tool that obtains hot news through its built-in data crawling layer, and intelligently filters high-quality topics that can be associated with investment and financial management. Suitable for financial content creation scenarios that need to guide securities account opening and fund investment.

6installs
Sourcecyhzzz/finance_aigc_skills
Added on

NPX Install

npx skill4agent add cyhzzz/finance_aigc_skills hot_topics_selector

Tags

Translated version includes tags in frontmatter
financial-topic-selectionai-agentdata-crawlinginvestment-guidancecontent-creation

SKILL.md Content (Chinese)

View Translation Comparison →

Financial Hot Topic Selection Tool

This Skill is used to filter high-quality topics suitable for guiding investment and financial management from financial hot spots.

Core Workflow

Phase 0: Data Crawling (Python)
   → Obtain 133 news headlines
Phase 1: Agent Intelligent Screening (Markdown Prompt)
   → Filter out 5 topics
Phase 2: Content Crawling (Python)
   → Crawl detailed content of 5 news articles
Phase 3: Agent Intelligent Parsing (Markdown Prompt)
   → Generate 5 complete topic plans
Separation of Responsibilities:
  • Python Scripts: Data Crawling (Phase 0, 2)
  • Agent: Intelligent Screening and Parsing (Phase 1, 3)

Phase 0: Data Crawling

Execution Method:
bash
cd scripts
python3 fetch_hot_topics.py
Output: 
/tmp/hot_topics.json
Included Content:
  • 133 news headlines (from 5 platforms)
  • Weibo, Baidu, Toutiao, Douyin, Cailianshe
Data Format:
json
{
  "fetch_time": "2026-02-13T19:36:24+08:00",
  "total_items": 133,
  "data": {
    "weibo": { "items": [...], "count": 30 },
    "baidu": { "items": [...], "count": 30 }
  }
}

Phase 1: Agent Intelligent Screening

Agent Execution Prompt:
You are a financial content topic selection expert. Please filter out 5 topics most suitable for guiding securities account opening/fund investment based on the following news headlines.

Headline List

{133 headlines read from /tmp/hot_topics.json}

Screening Criteria (Investment Relevance as Core)

Highest Priority (90-100 points)

Can directly link to investment products:
  • Money-making stories → AI funds/tech stocks/brokerage products
  • Investment cases → Gold ETF/asset allocation
  • Financial management topics → Fund fixed-investment/securities account opening
Examples:
  • "AI earns 2 million yuan per month" → Can recommend AI funds (95 points)
  • "Gold earns 1.96 million yuan" → Can recommend Gold ETF (92 points)

High Priority (80-89 points)

Can link to financial planning:
  • Salary topics → Wage income vs investment income
  • Savings topics → Saving money vs investment
  • Consumption topics → Financial planning
Example:
  • "Year-end bonus of 180 million yuan" → Can recommend fund fixed-investment (85 points)

Medium Priority (60-79 points)

Barely can link:
  • Education topics → Education funds
  • Pension topics → Pension investment
Example:
  • "Hengshui High School has changed" → Can link to education funds (70 points)

Not Selected (<60 points)

Cannot link to investment:
  • Pure entertainment gossip
  • Pure social news
  • Topics unrelated to finance

Output Format (JSON)

json
{
  "selected_indices": [1, 3, 5, 7, 9],
  "reasons": {
    "1": "✅ Can directly link: AI money-making → AI fund investment. Investment relevance 95 points, high public attention",
    "3": "✅ Can directly link: Gold case → Gold ETF. Investment relevance 92 points, strong data impact"
  },
  "investment_angles": {
    "1": "Can recommend: AI-themed funds, tech stocks, brokerage AI products",
    "3": "Can recommend: Gold ETF, brokerage gold products, asset allocation services"
  },
  "investment_relevance_score": {
    "1": 95,
    "3": 92
  }
}
Strict Requirements:
  1. Fixed selection of 5 headlines (no more, no less)
  2. Each headline has an investment relevance score ≥80 points
  3. Specify the recommended investment product directions
  4. Prioritize topics that can directly link to products (≥90 points)

Phase 2: Content Crawling

Execution Method:
bash
cd scripts
python3 fetch_news_content.py \
  --input /tmp/hot_topics.json \
  --indices 1,3,5,7,9 \
  --output /tmp/news_content.json
Input: 5 topic indices output from Phase 1
Output: 
/tmp/news_content.json
Included Content:
  • Complete content of 5 news articles
  • Headline, URL, platform, body text, keywords
Data Format:
json
{
  "fetch_time": "2026-02-13T21:00:00+08:00",
  "total_articles": 5,
  "articles": [
    {
      "index": 1,
      "title": "Hangzhou guy runs a 1-person company and earns 2 million yuan per month with AI",
      "url": "https://...",
      "platform": "Weibo",
      "content": "Complete news content...",
      "keywords": ["AI", "entrepreneurship", "earn 2 million per month"]
    }
  ]
}

Phase 3: Agent Intelligent Parsing

Agent Execution Prompt:
You are a financial content planning expert. Please generate a complete topic plan for each news article based on the following 5 news contents.

News List

{5 news articles read from /tmp/news_content.json}

Parsing Requirements

1. News Content Summary

  • Extract core information (100-150 words)
  • Retain key data and cases
  • Highlight investment-related information

2. Topic Selection Reasons

  • Explain investment relevance (80-100 points)
  • Analyze public attention
  • Predict viral potential

3. Guidance Strategy

  • Entry Point: How to start from the news topic
  • Turning Point: How to naturally shift to investment
  • Product Recommendation: Specific investment products to recommend
  • Script Examples: 2-3 guiding scripts
Example Guidance Strategy:
Entry Point: AI entrepreneurship earns 2 million per month, triggering the thought "Can I do it too?"
Turning Point: Ordinary people don't have the skills, but can share the dividends through investment
Product Recommendation: AI-themed funds, tech ETFs, brokerage AI products
Script Examples:
  "Want to participate in AI dividends? No need to have technical skills, investing in AI funds allows you to get a share"
  "Open an account to buy AI funds, ordinary people can also enjoy the dividends of AI development"

4. Original Link

  • Retain the complete news URL
  • Facilitate subsequent review and citation

Output Format (JSON)

json
{
  "topic_plans": [
    {
      "index": 1,
      "title": "Hangzhou guy runs a 1-person company and earns 2 million yuan per month with AI",
      "summary": "A Hangzhou entrepreneur operates a 1-person company through AI tools, with a monthly income of 2 million yuan. The main business is...",
      "selection_reason": {
        "investment_relevance": 95,
        "mass_attention": "High (AI + money-making topic)",
        "viral_potential": "Extremely strong (data impact of 2 million yuan monthly income)"
      },
      "guidance_strategy": {
        "entry_point": "AI entrepreneurship earns 2 million per month, triggering the thought 'Can I do it too?'",
        "turning_point": "Ordinary people don't have the skills, but can share AI dividends through investment",
        "product_recommendation": "AI-themed funds, tech ETFs, brokerage AI products",
        "script_examples": [
          "Want to participate in AI dividends? Investing in AI funds allows you to get a share",
          "Open an account to buy AI funds, ordinary people can also enjoy the dividends of AI development"
        ]
      },
      "source_url": "https://..."
    }
  ]
}
Strict Requirements:
  1. Each topic plan must have all 4 sections
  2. Guidance strategy must be specific and operable
  3. Product recommendations must be related to brokerages/funds
  4. Scripts must be natural and not forced

Usage Example

Complete Workflow

bash
# Step 1: Data Crawling
python3 scripts/fetch_hot_topics.py

# Step 2: Agent Intelligent Screening
# Agent reads the Phase 1 prompt in this file and executes it

# Step 3: Content Crawling
python3 scripts/fetch_news_content.py \
  --input /tmp/hot_topics.json \
  --indices 1,3,5,7,9

# Step 4: Agent Intelligent Parsing
# Agent reads the Phase 3 prompt in this file and executes it
Total Time Consumption: Approximately 15 minutes

Script Explanation

scripts/fetch_hot_topics.py

Function: Calls the TrendRadar API to crawl hot news headlines from 5 platforms
Output: 133 news headlines (JSON)

scripts/fetch_news_content.py

Function: Crawls detailed news content based on topic indices
Input: Topic indices (e.g., 1,3,5,7,9)
Output: Complete content of 5 news articles (JSON)

Reference Materials

references/Topic Selection Methodology.md

Includes topic selection methodologies from top financial new media influencers, including:
  • 5 core methodologies (pain point-driven, emotional resonance, data impact, story expression, practical orientation)
  • Topic formulas (number + conflict + result, pain point + solution, etc.)
  • Evaluation criteria (virality 40%, relevance 30%, practicality 20%, compliance 10%)
Agent can refer to this file to optimize topic selection strategies as needed.

Notes

  • Investment relevance is core: Each topic must be able to naturally shift to securities account opening or fund investment
  • Fixed output of 5 topics: No more, no less, ensure quality
  • Investment relevance ≥80 points: Topics with scores below 80 will be excluded
  • Python is responsible for data: Agent does not need to perform data crawling
  • Agent is responsible for intelligence: Screening and parsing are completed by Agent