market-research

In-Depth Research Skill

Analyze $ARGUMENTS to determine the research type and target, then execute the following command.

Execution Command

python3 /Users/kikuchihiroyuki/stock-skills/.claude/skills/market-research/scripts/run_research.py <command> <target>

Natural Language Routing

For natural language → skill determination, refer to .claude/rules/intent-routing.md .

Output by Research Type

stock (Stock Research)

• Basic Information + Valuation (yfinance)
• Latest News (yfinance)
• X Sentiment (Grok API)
• In-Depth Analysis: News, Earnings Materials, Analyst Views, Competitor Comparison (Grok API)

industry (Industry Research)

• Trends, Key Players, Growth Drivers, Risks, Regulatory Trends (Grok API)

market (Market Overview)

• Price Movements, Macro Factors, Sentiment, Notable Events, Sector Rotation (Grok API)

business (Business Model Analysis)

• Business Overview (What It Earns From)
• Business Segment Composition (Segment Name, Revenue Ratio, Overview)
• Revenue Model (Stock Type/Flow Type/Subscription, etc.)
• Competitive Advantage (Barriers to Entry, Brand, Technology, Moat)
• Key KPIs (Metrics Investors Should Watch)
• Growth Strategy (Medium-Term Management Plan, M&A, New Businesses)
• Business Risks (Structural Risks, Dependencies)

About the API

Grok API

• Use the Grok API only when the XAI_API_KEY environment variable is set
• When not set, generate the report using only yfinance data (for stock)

Two-Layer Structure

1. Layer 1 (yfinance): Always available (fundamentals, stock price data)
2. Layer 2 (Grok API): When XAI_API_KEY is set (in-depth analysis via X posts and Web search)

• industry / market / business require Layer 2. When not set, display that fact

API Status Summary (KIK-431)

Display the Grok API status at the end of each report:

Status	Display
Normal	✅ Normal
Not Set	🔑 Not Set — Available when XAI_API_KEY is set
Authentication Error	❌ Authentication Error (401) — Please check XAI_API_KEY
Rate Limit	⚠️ Rate Limit (429) — Wait a while and retry
Timeout	⏱️ Timeout — Check network connection
Other Errors	❌ Error — Check stderr for details

Output Supplement

After displaying the script output as-is, Claude should supplement the following:

For stock

• Check the consistency between fundamental data and Grok research
• Point out any divergence between value score and market sentiment
• Supplement with any additional context that affects investment decisions

In the case of industry

• Supplement with Japan-market-specific circumstances (regulatory environment, barriers to entry, etc.)
• Propose related stock screening (coordination with /screen-stocks)

In the case of market

• Estimate the impact on the portfolio (coordination with /stock-portfolio)
• Mention any similar past cases

In the case of business

• Consider the relationship between segment composition and stock valuation
• Sustainability of the revenue model (stock-based is stable, flow-based is highly sensitive to the economy, etc.)
• Check whether competitive advantage is reflected in actual financial indicators (ROE, profit margin, etc.)
• /stock-report Supplement consistency with fundamentals together with the results of

Execution example

# 銘柄リサーチ
python3 .../run_research.py stock 7203.T
python3 .../run_research.py stock AAPL

# 業界リサーチ
python3 .../run_research.py industry 半導体
python3 .../run_research.py industry "Electric Vehicles"

# マーケットリサーチ
python3 .../run_research.py market 日経平均
python3 .../run_research.py market "S&P500"

# ビジネスモデル分析
python3 .../run_research.py business 7751.T
python3 .../run_research.py business AAPL

Prior knowledge integration rules (KIK-466)

If the output of get_context.py contains the following, integrate it with the research results and respond:

• Relationship with held stocks: If there are portfolio holdings in the sector being researched, "there is an impact on the 7203.T you hold → health check recommended"
• Past research (SUPERSEDES): Compare with the previous research on the same subject. "Previous (2 weeks ago): sentiment neutral → this time: somewhat bullish"
• Investment notes: If there are concerns or theses about the target stock, suggest updates by cross-checking with the research results
• Watchlist: If the research target is on the watchlist, add context such as "watched stock → material for buy-timing judgment"

Prompting recording of analysis conclusions

When a stock/business/industry research response contains a conclusion with a view or thesis on a stock:

💡 This analysis has not yet been recorded as an investment note. Would you like to record it as a thesis/concern?