US Macro News Monitor (Web Workflow)
Use this skill for US macro monitoring that can indirectly impact Vietnamese equities.
Data sources
Primary media sources:
- - Bloomberg: https://www.bloomberg.com/
- WSJ: https://www.wsj.com/
- Reuters Markets: https://www.reuters.com/markets/
Fallback (open/public macro sources):
- - Federal Reserve (FOMC/calendar/speeches): https://www.federalreserve.gov/
- BLS releases (CPI, jobs): https://www.bls.gov/
- BEA releases (GDP, PCE): https://www.bea.gov/
- FRED calendars/series metadata: https://fred.stlouisfed.org/
Tooling assumption
- - OpenClaw web fetch + Brave API available.
- No Python scripts needed for this skill.
Execution workflow (ordered)
- 1. Collect fresh headlines from primary sources.
- Replace blocked/paywalled coverage with fallback public sources.
- Classify items into macro themes.
- Tag each item with tone (
risk-on, risk-off, neutral). - Map transmission channels to Vietnam sectors and representative tickers.
- Run quality gate and confidence scoring.
- Produce required output sections with source links.
Theme taxonomy
Classify each item into these macro themes:
- - inflation
- rates/yields
- growth/employment
- USD/FX
- oil/energy
- geopolitics/trade
Summarization horizon:
- - 24h changes
- 7d narrative drift
Data quality gate (required)
Run this gate before finalizing output:
- 1. Freshness: include only items with explicit publish/update time; report window in ICT (
UTC+7). - Coverage: target >= 3 distinct sources; if < 3, downgrade confidence.
- De-duplication: remove near-duplicate headlines by normalized title + URL domain.
- Evidence tags: mark each claim as
Fact or Inference. - Missingness log: list blocked/paywalled sources and what fallback replaced them.
Shared confidence rubric (required)
Apply this common standard:
- -
High: >= 3 distinct sources, <= 24h freshness for core items, and no single-source dominance over 60%. - INLINECODE7 : 2 sources or partial freshness gaps, but core themes still cross-confirmed.
- INLINECODE8 : 1 source only, stale coverage, or major paywall blocks without reliable fallback.
Always output confidence with:
- 1. Source count and source list.
- Time window used (ICT
UTC+7). - Main uncertainty driver (paywall, missing source, stale timing, or weak cross-confirmation).
Output format (required)
- - Section 1: Top headlines with source URLs.
- Section 2: Theme counts and tone distribution.
- Section 3: Vietnam sector spillover map.
- Section 4: Actionable watchlist (what to monitor next).
Watchlist mode (optional)
If the user provides an
ACTIVE_WATCHLIST (tickers), extend Sections 3–4:
- - Add a
Watchlist Impact Map that maps macro themes → transmission channel → each ticker. - Provide monitoring triggers (what would confirm/negate the impact) and a confidence tag.
- Never output absolute buy/sell instructions; this skill produces signals and watch items.
Trigger examples
- - "What US macro news overnight could move VN stocks today?"
- "Summarize Fed/CPI signals and their impact on Vietnam sectors."
- "Map US rates and USD narrative to Vietnamese equities."
Quality rules
- - Clearly mark inferred impacts as inference.
- Respect paywalls: use only accessible metadata when full text is unavailable.
- Avoid over-claiming causality.
- If source coverage is below threshold, explicitly set confidence to
Low.
美国宏观新闻监测器(网络工作流)
使用此技能进行可能间接影响越南股市的美国宏观监测。
数据来源
主要媒体来源:
- - 彭博社:https://www.bloomberg.com/
- 华尔街日报:https://www.wsj.com/
- 路透市场:https://www.reuters.com/markets/
备用来源(开放/公共宏观数据):
- - 美联储(FOMC/日历/讲话):https://www.federalreserve.gov/
- 劳工统计局发布(CPI、就业):https://www.bls.gov/
- 经济分析局发布(GDP、PCE):https://www.bea.gov/
- 圣路易斯联储日历/系列元数据:https://fred.stlouisfed.org/
工具假设
- - 可使用OpenClaw网络抓取 + Brave API。
- 此技能无需Python脚本。
执行工作流(按顺序)
- 1. 从主要来源收集最新头条新闻。
- 用备用公共来源替换被屏蔽/付费墙覆盖的内容。
- 将项目分类为宏观主题。
- 为每个项目标注基调(风险偏好、风险规避、中性)。
- 将传导渠道映射至越南行业及代表性股票代码。
- 运行质量关卡和置信度评分。
- 生成所需输出部分并附上来源链接。
主题分类法
将每个项目归类为以下宏观主题:
- - 通胀
- 利率/收益率
- 增长/就业
- 美元/外汇
- 石油/能源
- 地缘政治/贸易
总结时间范围:
数据质量关卡(必需)
在最终确定输出前运行此关卡:
- 1. 新鲜度:仅包含有明确发布/更新时间的内容;以ICT(UTC+7)报告时间窗口。
- 覆盖度:目标≥3个不同来源;若<3个,则降低置信度。
- 去重:通过标准化标题+URL域名移除近似重复的头条新闻。
- 证据标签:将每项声明标记为事实或推断。
- 缺失记录:列出被屏蔽/付费墙的来源及替代的备用来源。
共享置信度评分标准(必需)
应用此通用标准:
- - 高:≥3个不同来源,核心项目新鲜度≤24小时,且单一来源占比不超过60%。
- 中:2个来源或部分新鲜度缺口,但核心主题仍可交叉验证。
- 低:仅1个来源,内容过时,或主要付费墙无可靠备用来源。
始终输出置信度信息,包括:
- 1. 来源数量及来源列表。
- 使用的时间窗口(ICT UTC+7)。
- 主要不确定性驱动因素(付费墙、缺失来源、时间过时或交叉验证薄弱)。
输出格式(必需)
- - 第1部分:头条新闻及来源URL。
- 第2部分:主题计数和基调分布。
- 第3部分:越南行业溢出效应映射图。
- 第4部分:可操作观察清单(下一步需监测的内容)。
观察清单模式(可选)
若用户提供ACTIVE_WATCHLIST(股票代码),则扩展第3-4部分:
- - 添加观察清单影响映射图,将宏观主题→传导渠道→每个股票代码进行映射。
- 提供监测触发条件(可确认/否定影响的因素)及置信度标签。
- 绝不输出绝对的买入/卖出指令;此技能仅产生信号和观察项目。
触发示例
- - 昨晚哪些美国宏观新闻可能影响今日越南股市?
- 总结美联储/CPI信号及其对越南行业的影响。
- 将美国利率和美元叙事映射至越南股市。
质量标准
- - 明确将推断的影响标注为推断。
- 尊重付费墙:当无法获取全文时,仅使用可访问的元数据。
- 避免过度声称因果关系。
- 若来源覆盖度低于阈值,明确将置信度设为低。