
equity-research-skill
机构级个股投研报告 Agent Skill(九章结构 + 多方法估值交叉验证),适用于 Claude Code / Cowork / Codex 等 AI Agent 工具
Install with your AI
Paste into Claude Code, Cursor, or any agent — it reads the repo and wires the tool into your project.
Install and set up equity-research-skill (claude-skill project) into my current project. Found on https://claudeers.com/equity-research-skill Repo: https://github.com/rollingSirius/equity-research-skill Homepage/docs: — Detected install method: claude-skill → # copy this skill into .claude/skills/equity-research-skill/ Category: skills. Platforms: cli, desktop. Read the repo's README for exact setup and env vars, then install it and wire it into my project. Claudeers Health Verdict: active; community-verified: false. Confirm the source before running anything.
# copy the skill dir into your project: # .claude/skills/equity-research-skill/ (or ~/.claude/skills/equity-research-skill/ for all projects)
git clone https://github.com/rollingSirius/equity-research-skill
// compatibility
| Platforms | cli, desktop |
|---|---|
| Operating systems | — |
| AI compatibility | claude |
| License | MIT |
| Pricing | open-source |
| Language | Python |
Equity Research Skill 个股投研报告技能
一个可用于 Claude Code / Claude Desktop (Cowork) / Codex 及其他 AI Agent 工具的 Agent Skill:把"研究一下某只股票"这类请求,转化为一份事实准确、来源可溯、结论明确的机构级九章投研报告(中文输出)。
An Agent Skill that turns "research this stock for me" into an institutional-grade, nine-chapter equity research report (written in Chinese) — with sourced data, timestamps, and multi-method valuation cross-checks.
它能做什么 What it does
- 多市场覆盖:美股 / 港股 / A股,A/H 双重上市标的默认做溢价对比并分市场给结论。US, HK and China A-shares, with an A/H premium module for dual-listed names.
- 五步工作流:确认标的 → 探测可用数据源并并行采集(行情 / Morningstar / 财报与卖方观点,工具缺失自动降级并标注)→ 数据对账与时间戳 → 按九章模板撰写 → 估值交叉验证并交付文件。
- 九章报告结构:一页速览、业务详情、竞争与护城河、管理层与治理、财务分析、多方法估值、分析师观点汇总、新闻与催化剂、投资结论。
- 估值纪律:至少三种方法交叉验证——反向 DCF(证伪现价)+ 三情景概率加权 DCF + 相对估值/SOTP;所有 DCF 计算由
scripts/dcf.py执行,禁止心算,假设以 JSON 留档。 - 结论可复现:估值标签与买卖动作按预注册标定规则映射产出(含动作矩阵与治理否决项),同一组数字不会两次给出不同结论。
- 研究纪律:事实与判断分离、每个关键数据标注来源与时间、冲突数据显式对账、缺失数据如实标注"未获取到"。
文件结构 Structure
equity-research-skill/
├── SKILL.md # 技能主文件(触发条件 + 工作流程)
├── references/
│ ├── report-template.md # 九章报告模板与表格骨架
│ ├── data-sources.md # 取数手册:工具探测与降级表、行情 / Morningstar / SEC / 分析师数据与对账规则
│ ├── valuation-methods.md # 估值方法:相对估值、正向/反向 DCF、情景加权、SOTP + 结论标定规则
│ └── markets-cn-hk.md # A股/港股/A+H 差异手册:合约解析、披露源、财年、A/H 溢价模块
└── scripts/
└── dcf.py # DCF 计算器:三阶段/反向/敏感性/概率加权/标定(python dcf.py --demo 自检)
安装 Installation
Claude Code
# 个人级(所有项目可用)Personal (all projects)
git clone https://github.com/rollingSirius/equity-research-skill.git ~/.claude/skills/equity-research
# 或项目级(随仓库共享)Or project-level (shared via repo)
git clone https://github.com/rollingSirius/equity-research-skill.git .claude/skills/equity-research
Claude Desktop / Cowork
把本仓库打包为 zip(或下载 Release),在 Settings → Capabilities → Skills 中上传。 Zip this repo and upload it via Settings → Capabilities → Skills.
Codex / 其他 Agent 工具 (Codex CLI & other agents)
本技能是纯 Markdown 指令,无私有格式。任何能读文件的 Agent 都能用: This skill is plain Markdown — any file-reading agent can use it:
- 把本仓库放进项目目录(如
skills/equity-research/)。 - 在你的 Agent 配置(如 Codex 的
AGENTS.md)中加入一句:当用户要求研究/分析某只股票时,先读取并遵循
skills/equity-research/SKILL.md的完整流程。
使用 Usage
方式一:自然语言自动触发(Claude Code / Claude Desktop / Cowork) Auto-trigger — just ask naturally:
"帮我研究一下 NVDA" | "分析下 Marvell 值不值得买" | "Is AMD overvalued?"
只要出现「公司名/代码 + 投研意图」,技能就会按 SKILL.md 的描述自动触发。
方式二:显式指定技能 Explicit invocation per tool:
| 工具 Tool | 调用示例 Example |
|---|---|
| Claude Code | 斜杠命令:/equity-research 分析 NVDA;或说:"用 equity-research 技能研究 TSLA" |
| Claude Desktop / Cowork | 对话中说:"用 equity-research 技能帮我看看 AAPL 值不值得买" |
| Codex CLI | 无原生技能机制,配置 AGENTS.md(见上)后说:"按 equity-research 流程分析 NVDA";或直接说:"先读 skills/equity-research/SKILL.md,再按它分析 NVDA" |
| 其他 Agent | 提示词加一句:"先读取 skills/equity-research/SKILL.md 并严格按其流程执行,然后研究 <股票>" |
依赖 Dependencies
| 依赖 | 必需性 | 说明 |
|---|---|---|
| 联网搜索 / 网页抓取 | 必需 | 获取 Morningstar、财报申报(SEC/巨潮/披露易)、分析师评级、新闻 |
| Python 3 | 必需 | 运行 scripts/dcf.py(仅标准库,无第三方依赖) |
| Interactive Brokers (IBKR) MCP | 可选 | 实时行情快照与历史走势;未连接时按 data-sources.md 第 0 节降级表自动改用网络行情源 |
| Morningstar MCP | 可选 | 有则直取结构化字段,无则网页抓取 |
| docx 技能 | 可选 | 仅当需要输出 Word 版报告 |
免责声明 Disclaimer
本技能产出的内容仅为研究参考,不构成投资建议。作者与本技能均非持牌投资顾问,投资决策及其后果由使用者自行承担。
Reports generated by this skill are for research reference only and do not constitute investment advice. Use at your own risk.
License
// faq
What is equity-research-skill?
机构级个股投研报告 Agent Skill(九章结构 + 多方法估值交叉验证),适用于 Claude Code / Cowork / Codex 等 AI Agent 工具. It is open-source on GitHub.
Is equity-research-skill free to use?
equity-research-skill is open-source under the MIT license, so it is free to use.
What category does equity-research-skill belong to?
equity-research-skill is listed under skills in the Claudeers registry of Claude-compatible tools.
// embed badge
[](https://claudeers.com/equity-research-skill)
// retro hit counter
[](https://claudeers.com/equity-research-skill)
// reviews
// guestbook
// related in Claude Skills
An agentic skills framework & software development methodology that works.
💫 Toolkit to help you get started with Spec-Driven Development
AI coding assistant skill (Claude Code, Codex, OpenCode, Cursor, Gemini CLI, and more). Turn any folder of code, SQL schemas, R scripts, shell scripts, docs,…