Serenity Interactive Chain Skill
English version
与其他仓库不同，Serenity Interactive Chain 使用 Serenity 的逻辑输出一个自包含、可交互的 HTML 页面（默认；也可按需生成 Cursor .canvas.tsx），能直接在浏览器打开、文字可选中/复制/搜索，并在与 AI 对话中持续完善整个投资地图。
它的核心不是“直接报股票”，而是先看真实的物理系统：业务怎么运转、钱流向哪里、哪一层最难扩产、证据强不强、市场有没有定价，最后才映射到公司。
这个 skill 强制走更慢但更干净的路径：
系统变化 -> 业务域拆解 -> 供应链层级 -> 卡点节点
-> 证据等级 -> 市场状态 -> 公司映射 -> 验证路径
它想回答的问题是：
哪个小而关键的环节，可能卡住更大的下游系统？
核心能力
- 先把复杂公司或生态拆成业务域。
- 再把每个业务域拆成供应链层级。
- 默认生成自包含 HTML 交互页面（脑图导航 + 节点详情 + 底部个股分析），也可按需生成 Cursor canvas。
- 支持从任意节点继续递归下钻。
- 每个节点都有标签：
- 每个节点可以继续触发动作：
- 底部「个股分析详情」可搜索、可逐条收起（含全部展开/收起），每只股票卡片带
重新评估 / 下钻 / 交叉验证，节点里的股票行有 查看分析 跳转。
- 在 HTML 模式下，动作按钮会复制带文件路径与节点/股票信息的提示词（浏览器无法直接调用 Cursor）。
- 既适合 Cursor Canvas，也能在 Claude Code / Codex 这类文本环境中使用。
HTML 示例
什么时候使用
适合这类需求：
用 serenity-interactive-chain 拆一下 Musk ecosystem，做成可交互脑图。
拆解 SpaceX 产业链，但要先分 Starlink、Starship、Starshield、地面基础设施和算力。
做一张 AI 数据中心电力供应链卡点图，找还没被充分定价的环节。
把 D5-N3-power-megapack 继续下钻，并更新当前 canvas 文件。
重新评估行星滚柱丝杠是不是仍然是当下卡点，并补 A股映射。
安装
克隆仓库：
git clone https://github.com/submato/serenity-interactive-chain-skill.git
cd serenity-interactive-chain-skill
安装到 Agent skill 目录：
mkdir -p ~/.agents/skills/serenity-interactive-chain
cp SKILL.md ~/.agents/skills/serenity-interactive-chain/SKILL.md
安装到 Claude Code：
mkdir -p ~/.claude/skills/serenity-interactive-chain
cp SKILL.md ~/.claude/skills/serenity-interactive-chain/SKILL.md
安装到 Codex 风格 skill 目录：
mkdir -p ~/.codex/skills/serenity-interactive-chain
cp SKILL.md ~/.codex/skills/serenity-interactive-chain/SKILL.md
免责声明
本项目仅用于信息跟踪和研究辅助。
它不提供投资建议、交易指令、个性化金融建议或收益承诺。所有财报、合同、价格、客户关系、监管约束和市场状态都应由使用者自行核实。
English
Serenity Interactive Chain is an Agent Skill for turning companies, industries, and ecosystems into interactive supply-chain chokepoint maps.
Unlike one-shot research outputs, this skill uses the Serenity-style supply-chain logic to generate a self-contained interactive HTML page by default (a Cursor .canvas.tsx only on request). It opens in any browser with selectable/copyable/searchable text, and can be refined through follow-up conversations with an AI agent, node by node, until the whole investment map becomes sharper.
It deliberately avoids jumping straight from a narrative to tickers. Instead, it follows a slower and cleaner path:
system change -> business-domain breakdown -> supply-chain layers -> chokepoint nodes
-> evidence grade -> market status -> company mapping -> verification path
The core question is:
Which small but critical layer could physically constrain a much larger downstream system?
Core Features
- Splits complex companies or ecosystems into business domains first.
- Maps each domain into supply-chain layers.
- Defaults to a self-contained interactive HTML page (mindmap nav + node detail + bottom stock-analysis), with Cursor canvas on request.
- Supports recursive drilldown from any node.
- Labels every node by:
- node type
- market status
- evidence grade
- chokepoint state
- Each node can trigger follow-up actions:
- generate sublayers
- add evidence
- map local-market stocks
- re-evaluate the node
- A bottom “stock analysis” section is searchable and per-card collapsible (with expand/collapse all); each stock card has
re-evaluate / drill down / cross-validate, and stock rows have a view analysis jump button.
- In HTML mode, action buttons copy a ready-to-paste prompt (with file path and node/stock info) since the browser cannot call Cursor directly.
- Works in Cursor Canvas and text-first environments such as Claude Code or Codex-style agents.
HTML Examples
When To Use
Use it for prompts like:
Use serenity-interactive-chain to map the Musk ecosystem as an interactive mindmap.
Break down the SpaceX supply chain, but split Starlink, Starship, Starshield, ground infrastructure, and compute first.
Create an AI data-center power supply-chain chokepoint map and identify underpriced layers.
Deepen node D5-N3-power-megapack and update the current canvas file.
Re-evaluate whether the planetary roller screw node is still a current chokepoint and add A-share mappings.
Installation
Clone the repository:
git clone https://github.com/submato/serenity-interactive-chain-skill.git
cd serenity-interactive-chain-skill
Install to an Agent skill folder:
mkdir -p ~/.agents/skills/serenity-interactive-chain
cp SKILL.md ~/.agents/skills/serenity-interactive-chain/SKILL.md
Install for Claude Code:
mkdir -p ~/.claude/skills/serenity-interactive-chain
cp SKILL.md ~/.claude/skills/serenity-interactive-chain/SKILL.md
Install for Codex-style skill folders:
mkdir -p ~/.codex/skills/serenity-interactive-chain
cp SKILL.md ~/.codex/skills/serenity-interactive-chain/SKILL.md
Disclaimer
This project is for information tracking and research support only.
It does not provide investment advice, trading instructions, personalized financial advice, or guaranteed outcomes. Always verify current filings, contracts, market prices, customer relationships, regulatory constraints, and market status independently.
评论