金融机制分析_finance-core-analysis

以下为本文档的中文说明核心金融分析技能是一个面向深度金融机制分析的创作工具处于每日金融管道的第二阶段daily-finance 提供事实基础finance-core-analysis 进行机制分析最终由 finance-explosive-article 转化为爆款文章。它的核心功能是从第一阶段的每日金融简报中生成深度机制分析报告涵盖流动性分析、利率研究、风险偏好评估、资本流动追踪、政策解读和资产负债表约束等多个维度。使用场景主要是当用户需要基于 daily-finance 输出进行更深入的微信公众号风格分析时、或需要进行核心宏观和市场机制分析时。该技能要求以阶段一的文件作为事实基础通过外部数据源验证和更新关键信息然后生成独立的、可发布的 Markdown 分析文档。核心特点在于机制导向的分析方法。与表层新闻摘要不同该技能追求的是理解市场背后的驱动力不是发生了什么而是为什么发生以及还会怎样发展。设计原则强调分析的可操作性和可链接性结果既是独立可发布内容也可作为下游爆款文章的生产原料。该技能的价值不仅体现在直接的功能实现上还在于它与现有技术生态系统的良好兼容性和集成能力。无论是作为独立工具使用还是嵌入到更大的工作流中它都能发挥出应有的作用Core Financial AnalysisOverviewGenerate adeep mechanism analysisfrom the stage-1 daily finance brief.This is stage 2 of the daily finance pipeline:daily-finance → finance-core-analysis → finance-explosive-articleUse the stage-1 file as the factual base, verify/update key data from current external sources, then produce a standalone publishable markdown analysis that can also feed the final 德哥风格爆款文章.InputPrefer reading:markdown/daily-finance-YYYY-MM-DD.mdIf the file is not provided:Use the newest matching file by filename/date only when the user clearly asks for the latest report; state this assumption in the output metadataAsk which report date or file to use when multiple plausible dates exist and “latest” is not impliedIf the user gives raw daily-finance content in chat, use that contentDo not invent missing factsExtract a reusable fact table before analysis:FieldWhat to captureReport dateDate used for file naming and source alignmentKey facts3-5 source-backed facts from daily-financeNumbersActual, expected/previous, unit, direction, close/intraday statusSourcesInherited source names/links and any reliability limitsOpen questionsFacts that require current re-check or must be excludedExternal data stepAlways access current external data when producing a current analysis.Use web search, browser, finance tools, or available MCP tools to:Re-check major market prices, yields, exchange rates, commodities, and volatility indicatorsVerify policy statements, macro data, earnings data, and geopolitical factsUpdate stale numbers from the stage-1 brief when newer reliable data existsAdd only source-verifiable dataUse source hierarchy:Primary: official releases, central banks, exchanges, regulators, company filings, Reuters, Bloomberg, FT, WSJChinese reliable sources: 财新, 第一财经, 财经, 21世纪经济报道, 经济观察报Secondary sources require backtracking to original dataCurrent-data pull should be narrow and decision-relevant. Prioritize:Rates: US 10Y/2Y, China 10Y, major central-bank signalLiquidity: DXY, CNH, SHIBOR/DR007 or equivalent local liquidity indicator when relevantRisk appetite: major equity index, VIX or local volatility/breadth proxyCommodities: oil, gold, copper only when linked to the day’s thesisIf a value cannot be verified, remove it or mark it as【待】; never build the core judgment on【待】.Core modelExplain market behavior through these variables:LiquidityInterest ratesRisk appetiteCapital flowsPolicy directionBalance-sheet pressureIncentive constraintsUse first-principles reasoning:事件 → 约束变化 → 资金行为 → 定价结果 → 风险信号Analysis rulesSeparate confirmed facts from interpretationReuse source-backed numbers from daily-financeAdd new data only when it is necessary and source-verifiableMark uncertainty clearlyWrite as a serious publishable 公众号 deep-analysis article; save the strongest viral packaging for finance-explosive-articleDo not give explicit buy/sell recommendationsFor every major conclusion, name the constraint that changed and the observable signal that would prove it wrongAvoid analyzing every lens mechanically; select the 3-5 lenses that actually explain the fact baseValidation checklistBefore final output, check:Data correctness: dates, units, directions, actual vs expected, intraday vs closeSource consistency: important claims have reliable sourcesLogic integrity: each conclusion follows from a mechanism, not from mood wordsCausal chain: event, constraint, capital behavior, pricing result, risk signalCounter-case: include what would make the judgment wrongPublication readiness: title, subheadings, short paragraphs, clear conclusionRequired structure1. TitleUse a clear 公众号-style titlePrefer tension and mechanism over clickbait2. Executive judgmentState the single most important market judgmentExplain what changed todayInclude one “what would change my mind” sentence3. Fact baseSummarize the 3-5 key facts inherited from daily-financePreserve important numbers and source labelsAdd newly verified external data when necessaryUse a compact table with columns:标签,事实,数值,来源,状态4. Mechanism analysisAnalyze through 3-5 lenses as relevant:LiquidityInterest ratesRisk appetiteCapital flowsPolicyBalance sheetsFor each lens, explain:What changedWhy it mattersHow it transmits into asset pricesWhat data would confirm or falsify this lens5. Core contradictionIdentify the main tension, for example:Growth vs inflationPolicy easing vs currency pressureRisk appetite vs earnings pressureLiquidity repair vs balance-sheet contraction6. Scenario deductionProvide:Base caseAlternative caseFalsification signalEach scenario must specify:Trigger conditionAsset-pricing implicationObservable confirmation signal7. Key variables to watchList 3-6 observable indicators for the next update.8. Non-advisory implicationExplain directional exposure and risk, not ticker calls.9. SourcesList inherited sources and any new verified sources.10. Disclaimer本文仅供参考不构成投资建议。File outputIf environment allows:Save to:markdown/finance-core-analysis-YYYY-MM-DD.mdRules:Use the same report date as the inputdaily-financefileCreate directory if missingUse UTF-8 encodingOutput a complete publishable markdown article, not notesIf write fails → fallback to chat outputWriting styleUse 公众号-readable structure: strong title, short paragraphs, numbered sectionsKeep the tone sharp but rationalExplain mechanisms in plain ChineseUse contrast where helpful: “不是A而是B”Avoid empty emotional phrases and unexplained jargonGoalProduce a reusable deep-analysis document that answers:What changedWhy it changedWhat mechanism connects facts to asset pricingWhat would prove the analysis wrong