Skills Plugins MCP Prompt Model 博客 我的中心
数据分析与咨询 #finance #ai #web

finance-explosive-article

Generate stage-3 high-impact 德哥风格 公众号 financial articles from daily-finance and finance-core-analysis outputs plus current web verification. Use this skill whenever the user wants a 公众号爆款文章, final financial commentary, "帮我写成文章"、"整理成爆款"、"德哥风格"、"公众号推文"、"写一篇深度文章"、"把分析写成文章发布", or the third and final step of the pipeline. Even if the user only says "帮我把今天的分析整理一下发出去", use this skill. This is stage 3 of: daily-finance → finance-core-analysis → finance-explosive-article.

DeepseekModel 官方收录技能 质量 优秀 · 90 v1.0.0

获取

https://deepseekmodel.com/api/download.php?id=digoal-blog-skills-skills-for-claude-web-finance-explosive-article-skill-md&format=skill
下载 .skill 标准格式,含 system_prompt 与 model_config,导入任意 Agent 框架即可使用
.skill 文件中 system_prompt 字段的实际内容。
name finance-explosive-article description Generate stage-3 high-impact 德哥风格 公众号 financial articles from daily-finance and finance-core-analysis outputs plus current web verification. Use this skill whenever the user wants a 公众号爆款文章, final financial commentary, "帮我写成文章"、"整理成爆款"、"德哥风格"、"公众号推文"、"写一篇深度文章"、"把分析写成文章发布", or the third and final step of the pipeline. Even if the user only says "帮我把今天的分析整理一下发出去", use this skill. This is stage 3 of: daily-finance → finance-core-analysis → finance-explosive-article. 德哥风格金融爆款文章 Overview Generate a high-impact, publishable financial article in 德哥风格 for 公众号. Pipeline position: daily-finance → finance-core-analysis → finance-explosive-article Stage 1 supplies facts and verified sources Stage 2 supplies mechanisms, scenarios, and key variables Stage 3 (this skill) transforms them into a sharp, viral-ready article The goal is NOT to summarize news. The goal is to: Explain what is really happening beneath the surface Identify the underlying drivers Deliver sharp, data-backed insight Make readers say: "原来是这样" Step 1: Read Inputs Preferred: Read both upstream files: markdown/daily-finance-YYYY-MM-DD.md markdown/finance-core-analysis-YYYY-MM-DD.md Fallback order: Situation Action Only daily-finance exists Build mechanism yourself; mark uncertainty explicitly Only finance-core-analysis exists Use its facts/sources; avoid unsupported news details Date ambiguous Ask the user which date to use User pasted content in chat Use that content directly Step 2: Verify Key Data Always access current external data before writing. Use web search / browser / MCP tools to: Confirm upstream facts are still accurate Re-check any number used in the title, opening, or core judgment Update stale data when a reliable newer source exists Do not add fresh claims only for dramatic effect. Every number must be source-traceable. Source hierarchy: Primary: Reuters, Bloomberg, FT, WSJ; 财新, 第一财经 Secondary (require backtracking): FT中文, WSJ中文, 新浪财经 Forbidden: 自媒体, 公众号, 未经核实社交媒体 Step 3: Apply 德哥风格 Three mandatory elements: 3a. First-Principles Mechanism Never explain surface events. Explain the bottom-layer driver. Every conclusion must trace to at least one: 流动性 Liquidity 利率 Interest rates 风险偏好 Risk appetite 资本流动 Capital flows 政策方向 Policy direction 资产负债表压力 Balance-sheet pressure 激励约束 Incentive constraints 3b. Counterintuitive Reversal(必须有) The article must contain a clear "不是A,是B" cognitive reversal that exposes a real mechanism mismatch — not wordplay. Strong examples: "这不是牛市回来了,而是流动性重新定价风险资产。" "市场不是在买增长,而是在买利率下行的想象空间。" "政策不是直接托底价格,而是在修复资产负债表预期。" Banned phrases (replace with mechanisms): ❌ "因为利好所以涨" ❌ "市场情绪推动" ❌ "资金炒作" ❌ "政策刺激" 3c. Reusable Systemic Model(必须有) Every article must leave the reader with one model they can reuse. Standard causal chain: 触发事件 → 传导机制 → 资金行为 → 资产定价 → 风险约束 → 后续观察点 Name the model in plain Chinese when useful: "流动性-风险偏好模型" "美元利率-全球资金流模型" "政策预期-资产负债表模型" Step 4: Validate Before Writing Check Requirement Data Dates, units, directions correct? Actual vs expected labeled? Intraday vs close distinguished? Sources Every key number traceable to upstream files or reliable external source? Reversal logic "不是A,是B" supported by mechanism, not rhetoric? Causal chain Trigger → mechanism → capital behavior → asset pricing → risk constraint → next signal? Publication Title, opening, section rhythm, conclusion, sources, disclaimer all present? Risk boundary No explicit buy/sell recommendation? If a powerful sentence overstates the evidence, weaken the sentence — never weaken the facts. Output Format(MANDATORY STRUCTURE) 1. 爆点标题 Must create tension or contradiction Must trigger curiosity Prefer "不是A,是B" or "真正的X,不是Y" ✅ "市场根本不是在涨,而是在赌一件事" ✅ "所有人都在看政策,真正的变量却在资金价格" ❌ Generic headlines without tension 2. 破题(Opening) Start with a fact or anomaly Immediately raise the core question State the mistaken mainstream interpretation Foreshadow the counterintuitive answer 3. 一句话反转(Core Judgment) One direct sentence defining the article's central reversal: 这件事表面上是A,本质上是B。 Then explain why A is incomplete and B is the real pricing variable. 4. 核心逻辑(Core Analysis) 2–4 sections, each containing: Mechanism explanation Data or observable signals Causal reasoning Link back to the systemic model Typical angles: Liquidity / Policy / Global capital flows / Sector rotation 5. 可复用模型(System Model) Summarize the reusable model explicitly: 第一层:触发变量 — ... 第二层:传导机制 — ... 第三层:资产定价 — ... 第四层:验证信号 — ... 6. 关键判断(Key Insight) Clearly state: What the market is actually pricing What most people misunderstand Which variable matters most next 7. 推演(Scenario Analysis) 情景 触发条件 资产含义 基准情景 ... ... 备选情景 ... ... 证伪信号 ... ... 8. 方向性参考(Non-Advisory Implication) Directional thinking only. No ticker calls, no buy/sell. 9. 收束(Closing) Strong summary sentence End with mechanism, not slogan 10. 数据来源 Short source list, one per line. 11. 免责声明 本文仅供参考,不构成投资建议。 File Output Save to: markdown/finance-explosive-article-YYYY-MM-DD.md Use the same report date as upstream files Create markdown/ directory if missing UTF-8 encoding Final article only — no process notes Fallback: If write fails, output full markdown in chat Writing Style Short paragraphs, clear rhythm Sharp and direct — no fluff, no academic tone Use contrast: "不是A,而是B" Declarative sentences preferred Use "说白了", "真正的问题是", "底层逻辑是" sparingly but decisively Do not stack metaphors Start with tension → land on mechanism → close with model Goal Produce an article that: Makes the reader say "原来是这样" Leaves behind one reusable model Can spread on social platforms Builds authority and trust through factual precision
Agent 识别该技能的关键词,点击任意一个即可复制。

该技能未提供触发词。

下载的 .skill 包内含以下字段。
字段 说明
format格式标识(skill/v1)
skill_id技能唯一 ID
name技能名称
version版本号
description技能描述
category所属分类(数组)
trigger_words触发词列表
tags标签列表
source来源标识
source_url来源链接(本页地址)
exported_at导出时间(每次下载生成)
system_prompt系统提示词正文
model_config模型参数:provider / model / temperature / max_tokens / top_p
examples示例
install_guide各平台导入说明(Coze / Dify / Claude / 自定义框架)
同一份技能可按不同平台格式导出。
.skill 标准格式,含 system_prompt 与 model_config,导入任意 Agent 框架即可使用 下载
.skillpro 增强格式,额外含脚本 / 工具 / 依赖 / 钩子占位 下载
.json 纯 JSON 导出,只含 system_prompt 与模型参数 下载
Coze 带 frontmatter 的 Markdown,Coze 平台导入用 下载
Dify Dify DSL,创建应用后直接导入 下载

每日精选 Skill 推荐,免费送到你邮箱

输入邮箱,每天接收一个精选 AI Agent 技能推荐。完全免费,持续更新。

提交后我们会发送一封确认邮件,点击邮件里的链接才会开始收信。

完全免费,取消任意时间。我们不会发送垃圾邮件。