功能描述

全球情报汇总工具 - 自动生成结构化的全球市场、政经、AI 新闻汇总报告。支持定向深度分析与智能推演。借鉴 situation-monitor 项目架构，增强 RSS 源接入、情报分级和高相关性事件检测。

使用说明 (SKILL.md)

全球情报汇总 Skill (v3.3)

Name: global-intel-summary
Author: gmsx000-cloud

自动生成结构化的全球情报汇总报告。v3.3 版本借鉴了 situation-monitor (GitHub 2.8k+ stars) 项目架构，新增了：

🆕 RSS 源聚合：接入 30+ 优质 RSS 源，覆盖政治、科技、金融、政府、AI、情报六大领域
🆕 情报分级 (Alert System)：自动识别重大地缘政治事件（战争、制裁、军事行动）
🆕 相关性检测：识别跨新闻的高相关性事件，识别"叙事演进"
🆕 区域热点：自动标记新闻涉及的关键地区（欧洲、中东、亚太、美洲、非洲）

一、执行要求（v3.3 核心铁律）

⚠️ 铁律 1：数据时效性

只使用 24 小时内最新数据
超过 48 小时的一律不用
如果没有最新数据，明确写"今日无新报道"

⚠️ 铁律 2：数据源强制

资产类别	强制数据源
加密货币	Binance（首选） + CoinMarketCap（验证）
美股	Yahoo Finance
A股/港股	东方财富/同花顺
国际政经	BBC/Reuters/X + RSS 聚合
地缘/情报	RSS 深度源（见下方列表）

⚠️ 铁律 3：强制溯源

每条数据必须标注：

时间戳（如：2026-02-14 09:30 UTC+8）
来源（如：来源: Binance | 时间: 09:25）

⚠️ 铁律 4：情报分级

以下关键词出现时，标记为 [⚠️ ALERT] 并提升优先级：

war, invasion, military, nuclear, sanctions, missile, attack, troops, conflict, strike, coup, martial law, emergency, nato

二、核心数据源 (v3.3 新增)

🆕 RSS 源矩阵（situation-monitor 推荐）

1️⃣ 政治/综合 (Politics)

BBC World: https://feeds.bbci.co.uk/news/world/rss.xml
NPR News: https://feeds.npr.org/1001/rss.xml
Guardian World: https://www.theguardian.com/world/rss
NYT World: https://rss.nytimes.com/services/xml/rss/nyt/World.xml

2️⃣ 科技 (Tech)

Hacker News: https://hnrss.org/frontpage
Ars Technica: https://feeds.arstechnica.com/arstechnica/technology-lab
The Verge: https://www.theverge.com/rss/index.xml
MIT Tech Review: https://www.technologyreview.com/feed/
ArXiv AI: https://rss.arxiv.org/rss/cs.AI

3️⃣ 金融 (Finance)

CNBC: https://www.cnbc.com/id/100003114/device/rss/rss.html
MarketWatch: https://feeds.marketwatch.com/marketwatch/topstories
Yahoo Finance: https://finance.yahoo.com/news/rssindex
FT: https://www.ft.com/rss/home

4️⃣ 政府/宏观 (Gov/Macro)

White House: https://www.whitehouse.gov/news/feed/
Federal Reserve: https://www.federalreserve.gov/feeds/press_all.xml
SEC Announcements: https://www.sec.gov/news/pressreleases.rss
DoD News: https://www.defense.gov/DesktopModules/ArticleCS/RSS.ashx?max=10&ContentType=1&Site=945

5️⃣ 情报/防务 (Intel/Defense)

CSIS: https://www.csis.org/analysis/feed
Brookings: https://www.brookings.edu/feed/
CFR: https://www.cfr.org/rss.xml
Defense One: https://www.defenseone.com/rss/all/
War on Rocks: https://warontherocks.com/feed/
Breaking Defense: https://breakingdefense.com/feed/
The Drive War Zone: https://www.thedrive.com/the-war-zone/feed

6️⃣ 区域深度 (Regional)

The Diplomat (亚太): https://thediplomat.com/feed/
Al-Monitor (中东): https://www.al-monitor.com/rss
Bellingcat (OSINT): https://www.bellingcat.com/feed/
CISA Alerts (网络安全): https://www.cisa.gov/uscert/ncas/alerts.xml

🆕 关键词检测规则

🚨 告警关键词 (Alert Keywords)

出现以下词汇时，标记为 [⚠️ ALERT]： war, invasion, military, nuclear, sanctions, missile, attack, troops, conflict, strike, coup, martial law, emergency, nato, ceasefire, treaty, assassination, terrorist

🌍 区域关键词 (Region Keywords)

欧洲 (EUROPE): nato, ukraine, russia, germany, france, uk, britain, poland
中东 (MENA): iran, israel, saudi, syria, iraq, gaza, lebanon, yemen, houthi
亚太 (APAC): china, taiwan, japan, korea, south china sea, asean, philippines
美洲 (AMERICAS): us, america, canada, mexico, brazil, venezuela
非洲 (AFRICA): africa, sahel, niger, sudan, ethiopia, somalia

三、执行步骤（v3.3）

Step 1: 获取当前时间

date

Step 2: 分批搜索（每批 ≤50 词）

搜索1：BTC/ETH/SOL Binance 价格
搜索2：美股指数（S&P 500, Nasdaq, Dow Jones）
搜索3：Trump/Musk 最新动态
搜索4：RSS 源深度搜索（根据当日热点选择 2-3 个 RSS 源关键词）
  - 例如：如果有中东冲突迹象，搜索 "Israel Iran Gaza latest"
  - 如果有 AI 突破，搜索 "OpenAI GPT-5 latest"
  - 如果有金融风险，搜索 "Fed rate decision latest"

Step 3: 情报分级处理

扫描所有新闻标题，检测 Alert Keywords
如果检测到 [⚠️ ALERT] 事件，在报告中优先展示，并标注 [⚠️]
标注涉及的 Region

Step 4: 数据校验

关键数据必须二次搜索验证
如果多源数据不一致，写"待核实"

Step 5: 生成报告

按模板输出，每条数据带时间戳 + 来源

四、报告输出模板 (v3.3)

# 🌍 全球情报汇总 (YYYY-MM-DD HH:MM)

---

## ⚠️ 今日热点 [ALERT]
*(如果有检测到 Alert 关键词的事件，放在这里)*

## 💰 加密货币
- BTC: $XX,XXX | 来源: Binance | 时间: HH:MM
- ETH: $X,XXX | 来源: Binance | 时间: HH:MM

## 📈 美股行情
- S&P 500: X,XXX.XX (X.XX%) | 来源: Yahoo Finance | 时间: HH:MM
- Nasdaq: XX,XXX.XX (X.XX%) | 来源: Yahoo Finance | 时间: HH:MM

## 🇨🇳 A股/港股
- 恒生指数: XX,XXX (X.XX%) | 来源: 东方财富 | 时间: HH:MM

## 🌍 国际政经
- [⚠️ ALERT] [区域] [标题] | 来源: XXX | 时间: HH:MM
- Trump: [动态] | 来源: X | 时间: HH:MM
- Musk: [动态] | 来源: X | 时间: HH:MM

## 🤖 AI 前沿
- [新闻1] | 来源: XXX | 时间: HH:MM
- [新闻2] | 来源: XXX | 时间: HH:MM
- [新闻3] | 来源: XXX | 时间: HH:MM

---

## 🔮 智能推演
基于今日情报，推演未来 48 小时的 3 个潜在走向：
1. [方向1] - 逻辑支撑
2. [方向2] - 逻辑支撑
3. [方向3] - 逻辑支撑

五、禁止行为

❌ 禁止使用超过 48 小时的数据
❌ 禁止编造数据（不知道就是不知道）
❌ 禁止用预测数据冒充实时数据
❌ 禁止使用非指定数据源

六、搜索技巧

加密货币搜索

"BTCUSDT Binance price today" (英文)
"比特币 价格 今日 Binance" (中文)

美股搜索

"S&P 500 Nasdaq Dow Jones February 14 2026"

RSS 源深度搜索

"site:feeds.bbci.co.uk middle east latest"
"site:csis.org defense latest"
"site:brookings.edu policy latest"

人物追踪

"Trump policy February 2026"
"Musk xAI AGI news February 2026"

七、进阶：相关性检测 (Narrative Tracking)

此为 v3.3 新增的高级分析功能

在生成报告时，尝试识别以下模式：

单点事件：孤立的重大新闻
叙事演进：多个相关新闻，描述同一事件的演进（如：事件A发生 -> 各方反应 -> 后续发展）
跨领域关联：科技 + 金融，科技 + 地缘政治的关联

提示：使用 LLM 的上下文理解能力，扫描当日新闻，识别这些模式。

由小爱开发 | v3.3 借鉴 situation-monitor 架构增强版

安全使用建议

Key things to consider before installing: - Mismatch between promise and implementation: The SKILL.md promises comprehensive RSS aggregation, alerting, and Binance-first crypto verification, but the only code provided (get_market_data.py) only queries CoinGecko and two Yahoo Finance endpoints and contains no RSS parsing or alert logic. Ask the author for the missing components or an explanation of how RSS aggregation and Binance verification will be performed. - Network access: The instructions expect the agent to perform broad web searches and fetch many external feeds. If you allow this skill to run, it may retrieve content from many third-party sites. Run it in a sandbox or with network controls if you want to limit exposure. - Credentials and rate limits: SKILL.md names services (Binance, CoinMarketCap, some government feeds) that sometimes require API keys or have rate limits. The package requests no credentials—clarify whether API keys are needed and why they are not declared. - Functional testing: Before relying on outputs for decision-making, test the skill in a safe environment to verify that it actually collects from the claimed sources, respects the 24-hour freshness rule, and correctly flags ALERT keywords. Ask for unit tests or logs demonstrating RSS fetches and alert detection. - Source and provenance: The skill's homepage and source are unknown. Request the upstream repository or provenance information (license, authorship, links to the referenced 'situation-monitor' repo) to increase trust. What would change this assessment: if the author provides the missing code that implements RSS ingestion, alerting, and Binance/CMC access (or documents why those sources are only optional), or if the skill were trimmed to match the actual implemented scope (market data only), the assessment would move toward 'benign'. Conversely, discovery of hidden endpoints, undeclared credential usage, or code that exfiltrates data would elevate the concern.

功能分析

Type: OpenClaw Skill Name: global-intel-summary Version: 1.0.0 The skill is classified as suspicious due to its broad and sensitive information-gathering capabilities, primarily outlined in SKILL.md. It instructs the AI agent to aggregate and deeply search over 30 RSS feeds from various domains, including numerous intelligence, defense, and geopolitical organizations (e.g., CSIS, Brookings, CFR, Bellingcat, CISA Alerts). The instructions also detail performing targeted web searches using `site:` specific queries on these sources. While the stated purpose is to generate intelligence summaries, the extensive nature of these sources and the directed search capabilities represent a significant potential for misuse if the agent were compromised via prompt injection, allowing an attacker to leverage these powerful reconnaissance features for unintended purposes. The `get_market_data.py` script, however, is benign, fetching data from standard public APIs.

能力评估

⚠ Purpose & Capability

The SKILL.md claims a full RSS aggregation, alerting, multi-source verification (Binance primary, CoinMarketCap verification) and narrative-tracking system. The shipped Python script only fetches CoinGecko and two Yahoo Finance endpoints (no Binance, no CoinMarketCap, no RSS fetching, no alerting logic). No environment credentials are requested even though SKILL.md mandates Binance as the primary crypto source. This is a substantive mismatch between claimed capability and actual code/assets.

⚠ Instruction Scope

Runtime instructions instruct the agent to run arbitrary web searches, query many external RSS feeds, perform multi-source verification, and use an LLM to infer narrative patterns. Those instructions are broad and require the agent to fetch content from many third-party endpoints (public news sites, feeds, exchanges). The skill does not include code to fetch/parse the listed RSS feeds or implement the described alert-detection pipeline; it relies on the agent (or unspecified tooling) to perform those network operations, which increases the potential for unexpected network access or inconsistent behavior.

✓ Install Mechanism

No install spec (instruction-only with a small helper script). Nothing is downloaded from arbitrary URLs and no package manager installs are declared, so there is no additional install-time code execution risk beyond the included files.

ℹ Credentials

The skill does not request any environment variables or credentials, which reduces immediate risk. However, SKILL.md mandates use of Binance and CoinMarketCap as primary/verification crypto sources but provides no guidance for API keys or authenticated access. That mismatch may force the agent to scrape public pages or attempt unauthenticated API access. Ask the author whether authenticated APIs are required and why no credentials are declared.

✓ Persistence & Privilege

The skill does not request 'always: true' and has no install-time persistence. It does not claim to modify other skills or system-wide settings. Autonomous invocation is allowed by default (normal) but is not combined here with elevated privileges.

版本历史

v1.0.0

global-intel-summary v3.3.0 introduces a major upgrade focused on real-time, structured global intelligence and news reporting. - Integrated 30+ high-quality RSS sources across politics, tech, finance, government, defense, and regional news. - Added an automated "Alert System" for classifying major geopolitical events (war, sanctions, military, etc.). - Enabled cross-news correlation detection and narrative evolution tracking. - Auto-tags news with relevant global regions (Europe, MENA, APAC, Americas, Africa). - Enforces strict data timeliness (24h freshness) and source attribution for every data point. - Enhanced financial/market tracking with mandatory authoritative sources for crypto, US, and China/HK stocks. - New structured report templates and intelligence escalation rules for clear, actionable summaries.

元数据

Slug global-intel-summary

版本 1.0.0

许可证 —

累计安装 0

当前安装数 0

历史版本数 1

常见问题

global-intel-summary 是什么？

全球情报汇总工具 - 自动生成结构化的全球市场、政经、AI 新闻汇总报告。支持定向深度分析与智能推演。借鉴 situation-monitor 项目架构，增强 RSS 源接入、情报分级和高相关性事件检测。它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 428 次。

如何安装 global-intel-summary？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install global-intel-summary」即可一键安装，无需额外配置。

global-intel-summary 是免费的吗？

是的，global-intel-summary 完全免费（开源免费），可自由下载、安装和使用。

global-intel-summary 支持哪些平台？

global-intel-summary 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。

谁开发了 global-intel-summary？

由 gmsx000-cloud（@gmsx000-cloud）开发并维护，当前版本 v1.0.0。

global-intel-summary