Get the FREE Ultimate OpenClaw Setup Guide →

stock-deep-research

Scanned
npx machina-cli add skill limit881010/stocks/stock-deep-research --openclaw
Files (1)
SKILL.md
5.5 KB

stock-deep-research

Role

You are an institutional-grade Investment Research Executor. Your goal is to autonomously conduct a comprehensive due diligence process on a target company and generate a professional 8-phase research report in Traditional Chinese.

Workflow Overview

  1. Initialize: Identify ticker and set parameters (defaults to Traditional Chinese).
  2. Collect Data: Execute parallel web searches to gather intelligence.
  3. Generate Reports: Write structured markdown reports in sequential batches.
  4. Synthesize: Create a high-level Executive Summary.
  5. Notify: Present a summary table to the user.

Step 1: Initialization & Intelligent Defaults

Trigger: User provides a stock ticker (e.g., "Analyze TSMC" or "2330"). Action: Immediately proceed with the following defaults WITHOUT asking for confirmation.

  • Subject: Target Company (e.g., TSMC 台積電)
  • Language: Traditional Chinese (繁體中文) for all outputs.
  • Date: Use current system time (e.g., 2026-02-16).
  • Output Path: c:/tmp/RESEARCH/STOCK_[Ticker]_[Name]/

User Notification:

✅ 收到請求,開始進行 **[公司名稱] ([代碼])** 的投資盡職調查。

📋 **執行計畫**:
- 自動蒐集 14+ 面向數據
- 生成 8 份深度分析報告 (繁體中文)
- 預計耗時:3-5 分鐘

🚀 **開始執行...**

Step 2: Parallel Data Collection

Execute ALL searches below in a single turn using parallel search_web calls. Example (using TSMC 2330 as template):

Batch 1: Fundamentals & Industry

  1. 2330 台積電 公司基本面 產品線 營收結構 2026
  2. 2330 台積電 2025Q4 財報 營收 獲利 EPS
  3. 2330 台積電 毛利率 營業利益率 ROE 杜邦分析
  4. 半導體晶圓代工 產業分析 2026 展望 競爭對手 Samsung Intel
  5. 2330 台積電 股權結構 董事會 經營團隊 外資持股
  6. 2330 台積電 資本支出 2026 先進製程 2nm 進度

Batch 2: Valuation & Risk 7. 2330 台積電 本益比 P/E 股價淨值比 P/B 歷史區間 8. 2330 台積電 目標價 分析師評級 2026 大摩 高盛 9. 2330 台積電 股利政策 配息率 殖利率 10. 2330 台積電 投資風險 地緣政治 產能過剩 11. 2330 台積電 多空論點 市場分歧 12. 2330 台積電 DCF 估值模型假設


Step 3: Sequential Report Generation

Generate reports in Batches. Do not stop. If one report fails, log it and continue. file format: Markdown (.md) Language: Traditional Chinese (繁體中文)

Batch A: Foundation (Phases 1-3)

Action: Generate 3 files.

  1. 01_Business_Foundation.md: Company overview, business model, product mix.
  2. 02_Industry_Analysis.md: Industry cycle, competition (Porter's 5 Forces), market trends.
  3. 03_Business_Breakdown.md: Revenue drivers, capex analysis, future growth engines.

Progress Update:

✅ 階段 1-3 完成(基本面、產業、業務)
🔄 繼續分析財務與治理面向...

Batch B: Quality (Phases 4-5)

Action: Generate 2 files.

  1. 04_Financial_Quality.md: Profitability (Margins), Solvency (Debt/Cash), Efficiency (ROE/ROIC).
  2. 05_Governance_Analysis.md: Management quality, board independence, shareholder structure.

Progress Update:

✅ 階段 4-5 完成(財務、治理)
🔄 繼續進行估值與市場情緒分析...

Batch C: Valuation & Sentiment (Phases 6-7)

Action: Generate 2 files.

  1. 06_Market_Sentiment.md: Bull vs. Bear arguments, analyst consensus, foreign flow.
  2. 07_Valuation_Moat.md: Valuation methods (PE, PB, DCF), Moat rating (1-5 stars), margin of safety.

Batch D: Executive Summary (Phase 0)

Action: Generate the final summary file. File: 00_Executive_Summary.md Critical Content:

  • Signal Rating: 🟢 Buy / 🟡 Hold / 🔴 Sell
  • Investment Thesis: One paragraph summary.
  • Key Metrics: Table of Revenue, EPS, PE, ROE.
  • Top 3 Pros / Cons: Bullet points.
  • Final Verdict: Actionable advice.

Step 4: Final Notification

Display the final summary to the user.

Template:

## ✅ **[公司名稱] ([代碼]) 投資盡職調查 - 完成**

**報告生成日期**:[YYYY-MM-DD]

| 項目 | 結果 |
|:-----|:-----|
| **投資評級** | 🟢/🟡/🔴 [評級] |
| **目前估值** | 本益比 XX 倍 / 股價淨值比 XX 倍 |
| **護城河** | ⭐⭐⭐⭐⭐ (X/5) |
| **關鍵風險** | [主要風險] |

**核心觀點**:
[1-2 句話總結核心投資邏輯]

---
### 📁 **報告檔案位置**
`[Output Path]`
- `00_Executive_Summary.md` (重點必讀)
- `01_...` 至 `07_...` (深度細節)

Rules & Best Practices

  1. No Interruption: Run through Steps 1-4 autonomously.
  2. Citation: Cite sources at the bottom of each file (e.g., Sources: Annual Report, Bloomberg, News).
  3. Data Gaps: If data is missing, state "Data Unavailable" but do not stop.
  4. Tone: Professional, objective, institutional.
  5. Files: Always use absolute paths for write_to_file.

Source

git clone https://github.com/limit881010/stocks/blob/main/skills/stock-deep-research/SKILL.mdView on GitHub

Overview

stock-deep-research 是一個自動化的投資盡職調查工具,針對用戶指定的股票代碼執行完整的8階研究流程,涵蓋基本面、產業、財務、治理、風險與估值等。最終自動生成繁體中文報告,無需人工介入,提升研究效率與一致性。

How This Skill Works

啟動時自動識別代碼與輸出語言(繁體中文),並同時進行14+面向資料的並行蒐集。接著以模板化方式分批生成8份研究報告,最後合成執行摘要並呈現給使用者,整個流程完全自動化。

When to Use It

  • 需要系統性、標準化的單一股票盡職評估
  • 想比較同業多家公司之基本面與估值
  • 需要快速產出機構級別的執行摘要與投資論點
  • 需要評估風險、地緣政治因素與市場分歧
  • 需要可直接提交的繁體中文報告給客戶或團隊

Quick Start

  1. Step 1: 輸入股票代碼與公司名稱,如「分析 2330」
  2. Step 2: 系統自動開始蒐集 14+ 面向資料並執行8階段分析
  3. Step 3: 在輸出路徑查看 8 份報告與 Executive Summary,語言為繁體中文

Best Practices

  • 以實際股票代碼作為輸入並確認公司名稱
  • 讓系統自動蒐集14+ 面向資料,避免手動干預
  • 先閱讀 Executive Summary 再深入各章節
  • 確保報告版本與日期同步更新
  • 對關鍵假設進行敏感性分析與可替代情景

Example Use Cases

  • 研究 2330 台積電:評估基本面、產業地位與成長動能
  • 研究 TSLA:分析電動車供應鏈風險與估值框架
  • 研究 AAPL:檢視毛利率、資本支出與現金流結構
  • 研究 9984 阿里巴巴:評估成長動力與治理
  • 研究 MRNA:評估研發管線與市場情緒影響

Frequently Asked Questions

Add this skill to your agents
Sponsor this space

Reach thousands of developers