News Sentiment Scan

舆情监控与情绪分析技能。扫描港股、美股、A股等公司公告、新闻报道、券商研报、社交媒体（微博、雪球等），去噪后进行情绪打分（-10至+10），输出情绪温度计与重大事件清单。触发场景：舆情监控、情绪分析、新闻情绪、社交媒体情绪、上市公司消息面分析、研报解读、"监控XX股票舆情"、"XX新闻情绪如何"。

Security Scan

The skill claims to aggregate and de-noise news from many sources (Weibo, Xueqiu, Google News, Twitter/X, company filings, research reports), but the shipped code only uses yfinance (Yahoo) for news — the declared capabilities and instructions do not match the actual implementation.

Purpose & Capability

SKILL.md and the description claim multi-channel collection (公司公告、券商研报、微博、雪球、Google News、Twitter/X 等) and de-noising, but the included script (scripts/sentiment_scan.py) obtains news via yfinance.Ticker only and processes titles; there is no code to fetch from Weibo, Xueqiu, Google News, Twitter/X, regulatory filing APIs, or broker research. The requested/required environment, binaries, and credentials are minimal/none — inconsistent with the claimed multi-source scraping.

Instruction Scope

SKILL.md instructs running the Python script and promises broad scraping/denoising behavior. The script's runtime behavior is limited to fetching news/info via yfinance and performing local text-based sentiment heuristics; it does not read system files, environment variables, or contact hidden endpoints beyond what yfinance uses. The instructions over-promise functionality (social media, research reports, filings, anti-bot/waterarmy filtering) that is not implemented in the code.

Install Mechanism

No install spec is provided (instruction-only + a local script). The Python script depends on standard pip packages (yfinance, pandas) and prints an error instructing the user to pip install them if missing. There are no downloads from arbitrary URLs or archive extraction in the manifest.

Credentials

The skill declares no required environment variables or credentials. That is proportionate to the shipped implementation (yfinance-based). However, SKILL.md claims access to platforms that often require credentials or special handling (Weibo, Xueqiu, Twitter/X), yet no credential fields or scraping instructions are present — this is an omission/inconsistency that may reflect incomplete implementation or misleading documentation.

Persistence & Privilege

The skill does not request persistent presence (always: false) and is user-invocable. It does not modify other skills or system configuration. There is no indication of elevated privileges or automatic always-on behavior.

What to consider before installing

This skill overstates its capabilities: the README promises multi-source scraping and de-noising (Weibo, Xueqiu, Google News, Twitter/X, filings, broker reports) but the included script only uses yfinance to fetch news and runs local keyword/pattern heuristics. Before installing or using: (1) inspect the code yourself or run it in an isolated environment to confirm what networks it contacts (yfinance/Yahoo is the only evident external source); (2) do not assume it collects Weibo/Xueqiu/broker reports—ask the author for the missing implementations or credentials if you need those sources; (3) if you plan to use results for trading, remember it's not financial advice and the heuristic scoring is simplistic; (4) monitor network traffic if you need to verify there are no hidden endpoints; and (5) consider sandboxing or code review before granting broader access or relying on it in production.

舆情监控与情绪分析技能，扫描多渠道信息源并进行情绪打分。

功能

• 多渠道信息采集：公司公告、新闻报道、券商研报、社交媒体（微博、雪球）
• 去噪处理：过滤广告、水军、无关信息
• 情绪打分：-10（极度负面）至 +10（极度正面）
• 事件提取：识别并量化每条重大事件的情绪贡献
• 输出格式：情绪温度计 + 重大事件清单

情绪评分标准

分数	情绪标签	说明
+8 ~ +10	🤩 极度乐观	重大利好，股价可能大幅上涨
+5 ~ +8	😊 偏正面	利好消息，持续关注
+2 ~ +5	🙂 轻微正面	小幅利好，保持观察
-2 ~ +2	😐 中性	无明显影响
-5 ~ -2	😟 轻微负面	小幅利空，注意风险
-8 ~ -5	😰 偏负面	利空消息，谨慎操作
-10 ~ -8	😱 极度悲观	重大利空，股价可能大幅下跌

使用方式

基本命令

python3 {SKILL_DIR}/scripts/sentiment_scan.py <stock_code> [days] [market]

参数说明：

• stock_code：股票代码
  • 港股：XXXX.HK（如 0700.HK）或 5位数字（如 00700）
  • 美股：字母代码（如 AAPL、TSLA）
  • A股：6位数字（如 002594、600519）
• days：监控天数，默认 7
• market：市场标识（可选，auto 自动识别）

示例：

# 监控比亚迪近7天舆情
python3 {SKILL_DIR}/scripts/sentiment_scan.py 002594 7

# 监控腾讯近30天舆情
python3 {SKILL_DIR}/scripts/sentiment_scan.py 0700.HK 30

# 监控苹果近7天舆情
python3 {SKILL_DIR}/scripts/sentiment_scan.py AAPL 7 us

输出格式

==================================================
📰 舆情监控报告：{股票名称}（{代码}）
==================================================

📅 监控周期：最近{N}天
🔍 信息来源：{来源列表}
⏰ 生成时间：{时间戳}

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

🌡️ 情绪温度计：{分数}（{情绪标签}）

   -10  ════════════●════════════  +10
        😱        😐        🤩
        └──────────┘
            {当前情绪位置}

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📋 重大事件清单：

1. [{情绪图标}] {事件标题}
   来源：{来源} | 时间：{日期}
   情绪贡献：{贡献分数} | 置信度：{置信度}%
   
2. [{情绪图标}] {事件标题}
   来源：{来源} | 时间：{日期}
   情绪贡献：{贡献分数} | 置信度：{置信度}%
   
...

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📊 情绪统计：
├── 正面事件：{N}条（{占比}%）
├── 中性事件：{N}条（{占比}%）
├── 负面事件：{N}条（{占比}%）
└── 平均情绪：{平均分}

💡 操作建议：{建议}

情绪来源权重

来源	权重	说明
公司公告	1.0	官方信源，最权威
券商研报	0.9	专业机构分析
主流媒体	0.8	财经媒体新闻
监管文件	0.9	交易所、证监会文件
社交媒体	0.6	微博、雪球等
论坛帖子	0.5	股吧、社区讨论

事件类型与情绪贡献

事件类型	基础分数	说明
业绩超预期	+5	营收/利润大幅增长
研报上调评级	+3	券商目标价上调
大股东增持	+3	管理层看好
政策利好	+4	行业政策支持
产品发布	+2	新产品上市
业绩不及预期	-4	营收/利润下降
研报下调评级	-3	券商目标价下调
大股东减持	-3	管理层不看好
政策利空	-4	行业政策收紧
监管调查	-5	证监会/交易所调查

数据来源

1. Yahoo Finance：公司新闻、分析师评级
2. Google News：财经媒体报道
3. Twitter/X：社交媒体讨论
4. 微博：中国社交媒体（需要网络访问）
5. 雪球：投资社区讨论

注意事项

• 情绪分析基于自然语言处理，可能存在误差
• 重大事件需结合基本面判断
• 本工具仅供参考，不构成投资建议
• 社交媒体信息可能被操纵，请谨慎判断