AgentSkillsCN

sentiment-analysis

汇总并分析某只股票的新闻情绪

SKILL.md
--- frontmatter
name: sentiment-analysis
description: Aggregate and analyze news sentiment for a stock symbol
allowed-tools: get_news_sentiment
metadata:
  domain: news
  complexity: basic

Sentiment Analysis Workflow

OBJECTIVE: Determine overall news sentiment and key themes for the symbol.

Step 1: Gather Recent News

  • •Use get_news_sentiment with max_results=10
  • •Focus on news from the past 7 days

Step 2: Classify Each Item

For each news item, classify as:

  • •POSITIVE: Bullish news (earnings beat, upgrades, product launches)
  • •NEGATIVE: Bearish news (misses, downgrades, lawsuits, delays)
  • •NEUTRAL: Informational without clear sentiment

Step 3: Calculate Aggregate Score

  • •Count: X positive, Y negative, Z neutral
  • •Score = (positive - negative) / total
  • •Score > 0.3: BULLISH sentiment
  • •Score < -0.3: BEARISH sentiment
  • •Otherwise: MIXED sentiment

Step 4: Identify Themes

  • •What topics appear repeatedly?
  • •Any developing narratives?
  • •Institutional vs retail focus?

Output Format

Sentiment Score: [+X.XX or -X.XX] -> [BULLISH/BEARISH/MIXED] Distribution: {positive} positive, {negative} negative, {neutral} neutral Key Themes:

  1. •[Theme 1 with example headline]
  2. •[Theme 2 with example headline] Dominant Narrative: [Summary of overall story]