{
  "markdown": "# Claude Code Stock Deep Research Agent\n\n**Investment Research Edition** - 专业股票投资尽调系统\n\n  ⚖️ 免责声明\n\n  本研究报告不构成投资建议或推荐。所有投资存在风险，包括本金损失。\n\n  重要提示:\n  1. 本报告仅供教育和信息用途\n  2. 部分数据需要通过官方渠道验证\n  3. 过往业绩不代表未来表现\n  4. 投资决策前请自行进行尽职调查\n  5. 建议咨询合格的财务顾问\n\n  ---\n  🎓 研究框架\n\n  本研究基于 Claude Code Deep Research 系统：\n  - 方法论: 8阶段股票投资尽调框架\n  - 智能体: 28个并行研究智能体\n  - 工具: WebSearch、WebFetch、综合分析\n  - 质量: 多空平衡、明确风险、数据验证\n\n  ---\n \n## Table of Contents\n\n1. [Features](#features)\n2. [Repo Structure](#repo-structure)\n3. [Quick Start](#quick-start)\n4. [Stock Investment Research](#stock-investment-research)\n5. [How It Works](#how-it-works)\n6. [Customization](#customization)\n7. [Credits & Acknowledgements](#credits--acknowledgements)\n8. [License](#license)\n\n---\n\n## Features\n\nThis repository contains **two specialized deep research frameworks** for Claude Code:\n\n### 1. 🎯 Stock Investment Research (股票投资尽调系统) ⭐ PRIMARY\n\nAn **8-phase investment due diligence framework** for analyzing publicly traded companies, inspired by professional investment research methodologies.\n\n**Key Capabilities**:\n- 📊 **Comprehensive Analysis**: Business model, industry dynamics, financial quality, governance, valuation\n- 🤖 **Multi-Agent Research**: ~28 parallel research agents working concurrently\n- 📈 **Investment Style Adaptation**: Value, growth, turnaround, dividend investing\n- 💰 **Valuation Models**: DCF, reverse DCF, relative valuation, scenario analysis\n- 🛡️ **Risk Assessment**: Bear case, black swans, monitoring checklist\n- ✅ **Quality Assurance**: Cross-validation (profit vs. cash flow, company vs. peers)\n- 📝 **Structured Output**: 20-file standardized due diligence report\n\n**Research Coverage**:\n- A-shares (A股) - 中国大陆股市\n- Hong Kong stocks (港股)\n- US stocks (美股)\n- Other global markets\n\n**Output**: Signal rating (🟢🟢🟢 Strong Buy / 🟡🟡🟡 Hold / 🔴🔴 Avoid) based on fundamental analysis\n\n### 2. 📚 General Deep Research (通用深度研究系统)\n\nA flexible **7-phase framework** for general research topics (business, technology, academic, etc.).\n\n---\n\n## Repo Structure\n\n| File/Folder | Purpose |\n|-------------|---------|\n| **CLAUDE.md** | Master instructions for Claude Code |\n| **.claude/skills/stock-question-refiner/** | Stock research question refinement skill |\n| **.claude/skills/stock-research-executor/** | 8-phase investment due diligence executor |\n| **.claude/commands/stock-research.md** | Main stock research command |\n| **.claude/skills/citation-validator/** | Citation verification skill |\n| **.claude/skills/got-controller/** | Graph of Thoughts controller |\n| **.claude/skills/synthesizer/** | Findings synthesis skill |\n| **STOCK_RESEARCH_IMPLEMENTATION_PLAN.md** | Stock research system design document |\n| **CLAUDE2.md** | Graph of Thoughts implementation details |\n| **PROJECT_UNDERSTANDING.md** | Architecture deep dive |\n| **IMPLEMENTATION_GUIDE.md** | User guide |\n\n---\n\n## Quick Start\n\n### Stock Research (股票投资尽调)\n\n```bash\n# Start Claude Code\nclaude\n\n# Set model (optional, but recommended)\n/model opus\n\n# Execute stock research\n/stock-research [股票代码或公司名称]\n\n# Examples:\n/stock-research 600519              # 贵州茅台 (A-share)\n/stock-research AAPL                # 苹果公司 (US)\n/stock-research 腾讯 00700.HK       # 腾讯控股 (HK)\n```\n\n**The system will**:\n1. Ask about your investment style (价值/成长/困境/红利), holding period, risk tolerance\n2. Deploy ~28 parallel research agents across 7 phases\n3. Generate comprehensive due diligence report in `RESEARCH/STOCK_[ticker]_[company]/`\n\n**Time**: 2-4 hours for standard due diligence\n\n### Output Example\n\n```\nRESEARCH/STOCK_600519_Kweichow_Moutai/\n├── 00_Executive_Summary.md        # 🟡🟡🟡 Hold / Fairly Valued\n├── 01_Business_Foundation.md      # Products, revenue, customers\n├── 02_Industry_Analysis.md        # Industry cycle, competition\n├── 03_Business_Breakdown.md       # Profit drivers, economics\n├── 04_Financial_Quality.md        # Cash flow, margins, red flags\n├── 05_Governance_Analysis.md      # Ownership, management\n├── 06_Market_Sentiment.md         # Bull/bear cases\n├── 07_Valuation_Moat.md           # Moat rating, valuation\n├── Financial_Data/                # Metrics, trends, peer comparison\n├── Valuation/                     # DCF, scenarios\n├── Risk_Monitoring/               # Bear case, monitoring checklist\n└── sources/                       # Citations with quality ratings\n```\n\n---\n\n## Stock Investment Research\n\n### 8-Phase Due Diligence Process\n\n| Phase | Focus | Output |\n|-------|-------|--------|\n| **1. Business Foundation** | 公司事实底座 | Products, revenue mix, customers, value chain, strategy |\n| **2. Industry Analysis** | 行业周期分析 | Cycle stage, supply-demand, competition, policy impacts |\n| **3. Business Breakdown** | 业务拆解 | Segments, profit engines, pricing power, economics |\n| **4. Financial Quality** | 财务质量 | Metrics trends, cash flow vs. earnings, red flags, peers |\n| **5. Governance Analysis** | 股权治理 | Ownership, management, capital allocation, ROIC |\n| **6. Market Sentiment** | 市场分歧 | Bull case, bear case, key debates, verification nodes |\n| **7. Valuation & Moat** | 估值护城河 | Moat rating (0-5), relative/absolute valuation, risks |\n| **8. Final Synthesis** | 综合报告 | Signal rating, thesis, monitoring checklist |\n\n### Investment Style Adaptation\n\nThe system adapts research approach based on investment style:\n\n| Style | Focus | Valuation Methods | Key Metrics |\n|-------|-------|-------------------|-------------|\n| **Value** (价值投资) | Intrinsic value, margin of safety | P/B, EV/EBITDA, DCF (conservative) | P/B, normalized earnings, FCF yield |\n| **Growth** (成长投资) | TAM, competitive positioning | PEG, DCF (aggressive), user models | Revenue growth, moat, TAM penetration |\n| **Turnaround** (困境反转) | Liquidity, solvency, catalysts | Liquidation value, option value | Debt ratios, cash runway, catalyst timeline |\n| **Dividend** (红利投资) | Dividend sustainability | DDM, FCF yield, yield comparison | Dividend yield, payout ratio, FCF coverage |\n\n### Quality Standards\n\n**Mandatory Cross-Validation**:\n- ✅ Profit vs. cash flow (OCF/NI ratio analysis)\n- ✅ Company vs. peer comparison (key ratios, margins)\n- ✅ Bear case analysis (3-5 key risks identified)\n\n**Citation Requirements**:\n- Every factual claim must include: Author, Date, Title, URL\n- Source quality rated (A-E scale)\n- A-grade sources: Annual reports, regulatory filings\n- B-grade: Industry reports, analyst research\n- C-grade: News articles, expert commentary\n\n**Signal Rating** (based on fundamentals only):\n- 🟢🟢🟢 **Strong Buy**: Significant margin of safety, strong moat, attractive valuation\n- 🟡🟡🟡 **Hold**: Fairly valued or limited margin of safety\n- 🔴🔴 **Avoid**: Overvalued, deteriorating fundamentals, excessive risks\n\n### Example Usage\n\n```\nUser: /stock-research 贵州茅台 600519\n\nClaude: 我可以帮您研究贵州茅台。让我先了解您的投资需求：\n\n       1. 投资风格：价值投资/成长投资/困境反转/红利投资？\n       2. 持有周期：短期(<6月)/中期(6-18月)/长期(1-3年以上)?\n       3. 关注重点：生意质量/财务质量/估值/公司治理 (选2-3个)\n       4. 风险偏好：保守/平衡/激进？\n\nUser: 价值投资，长期持有3年以上，关注：生意质量、公司治理、估值，风险保守\n\nClaude: [Generating research plan...]\n       [Deploying 28 parallel research agents...]\n       [Analyzing 5-10 years of financial data...]\n       [Comparing with Wuliangye, Yanghe, Fenjiu...]\n       [Assessing competitive moat...]\n       [DCF valuation with 3 scenarios...]\n       [Identifying bear case risks...]\n\nOutput: RESEARCH/STOCK_600519_Kweichow_Moutai/\n\n🟡🟡🟡 Hold / Fairly Valued\n\nInvestment Thesis:\n茅台是中国领先的白酒品牌，拥有强大的品牌护城河和定价能力。然而当前估值（P/E 32x）安全边际有限。建议继续持有现有仓位，但新投资应等待10-15%回调至¥1,750-1,850区间。\n\nKey Metrics:\n- Market Cap: ¥2.8T\n- P/E (TTM): 32x (above 5-year average)\n- Gross Margin: 91.2%\n- ROE: 31%\n- Moat Rating: 5/5 (Very Strong)\n\nTop 3 Reasons to Consider:\n1. Unassailable brand moat with 800-year heritage\n2. Exceptional margins (91% gross, 53% net)\n3. Strong cash generation (OCF/NI > 1.0)\n\nTop 3 Reasons to Avoid:\n1. Full valuation (P/E 32x, limited margin of safety)\n2. Regulatory risk (government scrutiny of luxury pricing)\n3. Competitive intensification (Wuliangwa narrowing gap)\n\nMonitoring Checklist:\n✅ Strengthen: Price pulls back 10-15% to ¥1,750-1,850\n❌ Exit: Price drops below ¥1,300 (-35%), net margin < 45%\n\n[Full report: 20 files, 127 sources, 50+ pages]\n```\n\n---\n\n## How It Works\n\n### Stock Research Workflow\n\n```\nUser: /stock-research [ticker]\n  ↓\nstock-question-refiner skill\n  - Asks: Investment style? Holding period? Focus areas? Risk tolerance?\n  ↓\nStructured Research Prompt (investment parameters, priorities, constraints)\n  ↓\nstock-research-executor skill\n  ├─ Phase 1: Business Foundation (4 parallel agents)\n  ├─ Phase 2: Industry Analysis (4 parallel agents)\n  ├─ Phase 3: Business Breakdown (4 parallel agents)\n  ├─ Phase 4: Financial Quality (4 parallel agents)\n  ├─ Phase 5: Governance Analysis (4 parallel agents)\n  ├─ Phase 6: Market Sentiment (4 parallel agents)\n  └─ Phase 7: Valuation & Moat (4 parallel agents)\n  ↓\ncitation-validator skill\n  - Verifies all claims have citations\n  - Rates source quality (A-E)\n  ↓\nComprehensive Investment Due Diligence Report\n  - Signal rating\n  - 8 phase reports\n  - Financial data tables\n  - Valuation analysis\n  - Risk monitoring checklist\n```\n\n### Key Innovations\n\n1. **Investment Style Adaptation**: Research approach tailored to value, growth, turnaround, or dividend investing\n2. **Parallel Multi-Agent Execution**: ~28 agents working concurrently for efficiency\n3. **Mandatory Cross-Validation**: Profit vs. cash flow, company vs. peers, bear case analysis\n4. **Structured Output**: Standardized 20-file report format\n5. **Quality Assurance**: A-E source quality rating, citation verification\n\n### General Research Workflow (Secondary)\n\n```\n[ Question ] → [ stock-question-refiner ]\n      ↓\n[ Structured Prompt ]\n      ↓\n[ research-executor ]\n      ├─ Planning (break into subtopics)\n      ├─ Multi-Agent Research (parallel)\n      ├─ Source Triangulation (A-E rating)\n      └─ Synthesis (combine findings)\n      ↓\n[ Citation Validation ]\n      ↓\n[ Research Report ]\n```\n\n---\n\n## Customization\n\n### Adapting Stock Research Parameters\n\nThe system automatically adapts based on:\n\n1. **Investment Style**:\n   - Value: Emphasize balance sheet, normalized earnings, margin of safety\n   - Growth: Emphasize TAM, competitive positioning, growth sustainability\n   - Turnaround: Emphasize liquidity, solvency, catalysts\n   - Dividend: Emphasize payout sustainability, FCF generation\n\n2. **Holding Period**:\n   - Short-term (<6 months): Focus on near-term catalysts, sentiment\n   - Medium-term (6-18 months): Balanced approach\n   - Long-term (1-3+ years): Emphasize business sustainability, moat, intrinsic value\n\n3. **Risk Tolerance**:\n   - Conservative: Add filters (debt limits, minimum profitability)\n   - Balanced: Standard risk checks\n   - Aggressive: Accept higher volatility, focus on upside scenarios\n\n### Customizing Output\n\nAdjust research parameters by answering the question-refiner's questions with your specific needs:\n\n- Geographic focus (China, US, global)\n- Timeframe (3 years, 5 years, 10 years of data)\n- Source preferences (annual reports only, include news, etc.)\n- Language (Chinese, English, or bilingual)\n- Valuation methods (DCF required? sum-of-parts?)\n\n---\n\n## Credits & Acknowledgements\n\n### Stock Research Framework\n- **8-Phase Methodology**: Based on professional investment due diligence best practices\n- **Graph of Thoughts Framework**: [SPCL, ETH Zürich](https://github.com/spcl/graph-of-thoughts) (MIT License)\n- **Quality Standards**: Inspired by institutional investment research processes\n\n### Core System\n- **Research Methodology**: Inspired by OpenAI and Google Gemini deep research playbooks\n- **Prompt Generation**: Eliminates need for external question-refinement tools\n- **Claude Code Integration**: Leverages native Skills and Commands capabilities\n\n### Development\n- **Developed by**: Ankit at [My Business Care Team (MyBCAT)](https://mybcat.com)\n- **Stock Research Edition**: December 2025\n- **License**: MIT License (see LICENSE file)\n\n---\n\n## License\n\nMIT License. See `LICENSE` file for full details.\n\n---\n\n## Important Disclaimer\n\n⚠️ **WARNING / 重要提示**:\n\n**Stock Research System**:\n- This system does **NOT** provide investment advice\n- This system does **NOT** predict stock prices or provide target prices\n- Signal ratings (🟢🟢🟢/🟡🟡🟡/🔴🔴) are based **ONLY on fundamental analysis** (business quality, financial health, valuation, competitive moat)\n- All investments involve risk, including the loss of principal\n- Past performance does not guarantee future results\n- Always conduct your own due diligence and consult with qualified financial advisors before making investment decisions\n\n**股票研究系统**：\n- 本系统**不构成投资建议**\n- 本系统**不预测股价**\n- 信号灯评级仅基于**基本面分析**（业务质量、财务健康、估值、护城河）\n- 所有投资均有风险，包括本金损失\n- 过往表现不代表未来结果\n- 请自行进行尽职调查并在做出投资决策前咨询合格的财务顾问\n- 提示词参考来源：https://mp.weixin.qq.com/s/EFT5S-cCeCnEIDZOD_DYbA\n\n---\n\n## Support\n\nFor detailed documentation:\n- See `CLAUDE.md` for Claude Code instructions\n- See `STOCK_RESEARCH_IMPLEMENTATION_PLAN.md` for system design\n- See `.claude/skills/*/` for skill-specific instructions \n",
  "bytes": 12998,
  "sha": "934e98cf58be1a6faed1bad2206214d7f50a21216c00dae959135a65b79f5f14",
  "repo_slug": "liangdabiao/claude-code-stock-deep-research-agent",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/skl_liangdabiao_claude_code_stock_deep_resea_90bc2b83/readme"
}