Parse $ARGUMENTS to determine region, preset, and sector, then execute the following command.
Execution Command
python3 /Users/kikuchihiroyuki/stock-skills/.claude/skills/screen-stocks/scripts/run_screen.py --region <region> --preset <preset> [--sector <sector>] [--theme <theme>] [--top <N>] [--mode <query|legacy>]
Natural Language Routing
For natural language → skill determination, refer to .claude/rules/intent-routing.md .
Available Region Codes (yfinance EquityQuery)
Major regions: jp, us, sg, th, my, id, ph, hk, kr, tw, cn, gb, de, fr, in, au, br, ca, etc. (about 60 regions)
Available Exchange Codes
| Exchange | Code |
|---|---|
| Tokyo Stock Exchange | JPX |
| NASDAQ | NMS |
| NYSE | NYQ |
| Singapore Exchange | SES |
| Stock Exchange of Thailand | SET |
| Bursa Malaysia | KLS |
| Indonesia Stock Exchange | JKT |
| Philippine Stock Exchange | PHS |
| Hong Kong Stock Exchange | HKG |
| Korea Exchange | KSC/KOE |
| Taiwan Stock Exchange | TAI |
Theme Screening (KIK-439)
--theme <テーマ> can be combined with any preset to perform screening focused on a specific theme.
trending/pullback/alpha Presets are --theme not supported.
| Theme Key | Description | Target Industries (Excerpt) |
|---|---|---|
ai |
AI/Semiconductors | Semiconductors, Software—Infrastructure, Electronic Components |
ev |
EV/Next-Generation Vehicles | Auto Manufacturers, Electrical Equipment & Parts |
cloud-saas |
Cloud/SaaS | Software—Application, Software—Infrastructure |
cybersecurity |
Cybersecurity | Software—Infrastructure (Security-Focused) |
biotech |
Biotech/Drug Discovery | Biotechnology, Drug Manufacturers |
renewable-energy |
Renewable Energy | Solar, Utilities—Renewable |
fintech |
Fintech | Software—Application (Finance-Focused), Capital Markets |
defense |
Defense/Space | Aerospace & Defense |
healthcare |
Healthcare/Medical Devices | Medical Devices, Health Information Services |
Theme definitions are managed in config/themes.yaml .
Trend Theme Auto-Detection (KIK-440)
--auto-theme is specified, trend themes are automatically detected via the Grok API (X/Web search), and existing screening is executed sequentially for each theme.
Pipeline
- Grok API detects 3–5 notable themes from X/Web (with confidence scores)
themes.yamland only valid themes are executed (unsupported themes are skipped with notification)- Execute screening with the specified preset for each theme
Constraints
--auto-themeと--themeare mutually exclusive (cannot be used simultaneously)trending/pullback/alphaCannot be used together with presetsXAI_API_KEYRequired (exits with an error if not set)
trending Differences from presets
--preset trending |
--auto-theme |
|
|---|---|---|
| Detection Target | Trending on XIndividual Stocks | Trend'sTheme/Sector |
| Granularity | Stock-level granularity | Theme-level granularity |
| Screening | Fundamentals Evaluation | Evaluated with any preset within the theme |
Execution Examples
# 日本のトレンドテーマで割安株をスクリーニング
python3 .../run_screen.py --region japan --preset value --auto-theme
# 米国のトレンドテーマで高成長株
python3 .../run_screen.py --region us --preset high-growth --auto-theme
# グローバルのトレンドテーマでデフォルトプリセット
python3 .../run_screen.py --preset value --auto-theme
Screening Mode
--mode query(Default): EquityQuery Method. Uses the yfinance EquityQuery API to directly search for stocks matching conditions without needing a stock list. Supports all regions. Fast.--mode legacy: Stock List Method. Uses the conventional ValueScreener. Fetches and evaluates a predefined list of stocks (Nikkei 225, S&P 500, etc.) one by one. Supports only japan/us/asean.--with-pullback: Add pullback filter. Applies technical pullback detection additionally to any preset (value, high-dividend, etc.).--preset pullback. It is not necessary to specify it together with . (The pullback preset takes precedence.) Output uses the same column format as Pullback mode.
Preset
value: Traditional value investing (low PER, low PBR, ROE >= 5%)high-dividend: High-dividend stocks (dividend yield 3% or higher)growth: Pure growth stocks (high ROE >= 15%, revenue growth >= 10%. No PER constraint. Selected by growth potential regardless of undervaluation)growth-value: Growth value (growth potential + undervaluation)deep-value: Deep value (very low PER/PBR)quality: Quality value (high ROE >= 15% + undervalued. A superset of value with the ROE threshold tightened from 5% to 15%. Limited to highly profitable undervalued stocks)pullback: Pullback buying type (enter on a temporary correction during an uptrend. A 3-stage pipeline of EquityQuery -> technical -> SR. Execution takes time)alpha: Alpha signal (integration of 3 axes: undervaluation + quality of change + pullback. A 4-stage pipeline of EquityQuery -> quality of change -> pullback detection -> 2-axis scoring. Execution takes time)trending: X trend stocks (discover trending stocks on X with the Grok API -> evaluate fundamentals with Yahoo Finance.--themeYou can narrow down themes with . Requires XAI_API_KEY)long-term: Long-term investment suitability (high ROE >= 15%, EPS growth >= 10%, dividend >= 2%, PER <= 25, PBR <= 3, market cap >= 100 billion. Searches for stable growth stocks suitable for long-term holding)shareholder-return: Shareholder return focus (ranked by total payout ratio of dividend yield + buyback yield. With stability assessment: ✅Stable/📈Increasing/⚠️Temporary/📉Declining)high-growth: High-growth stocks (profit not required, revenue growth rate >= 20%, most recent quarterly revenue growth >= 10%, PSR <= 20, gross margin >= 20%. Loss-making growth companies are also eligible. Uses PSR instead of PER to prevent bubbles) (KIK-432)small-cap-growth: Small-cap rapid-growth stocks (market cap <= 100 billion, revenue growth rate >= 20%, PSR <= 15, gross margin >= 20%. Ten-bagger candidates undiscovered by institutional investors. With automatic market cap adjustment by region. Risk ★★★★) (KIK-437)contrarian: Contrarian candidates (technically oversold × undervalued valuation × solid fundamentals. 3-axis 100-point scoring. The opposite of a value trap, detecting 'excessive market reaction') (KIK-504)momentum: Momentum surge stocks (4-axis scoring of RSI/MACD/volume surge/trend alignment. Detects stocks in an uptrend.--submodeYou can choose stable acceleration or surge with ) (KIK-506)
Momentum screening details (KIK-506)
3-stage momentum classification
| Classification | Condition | Description |
|---|---|---|
| 🟢 Accelerating (accelerating) | Score 40–69 | Stable upward momentum. No sense of overheating and high continuity |
| 🟡 Surge (surge) | Score 70–89 | Sharp price increase. Short-term attention is high, but beware of a correction |
| 🔴 Overheated | Score 90–100 | Overheated state. High returns can be expected, but there is correction risk |
--submode Parameters
--submode You can narrow down the classifications displayed.
| Value | Display target | Recommended use case |
|---|---|---|
stable(Default) |
🟢 Accelerating only | When you want to select stocks in a stable uptrend |
surge |
🟡 Surge + 🔴 Overheated | When you want to include short-term surging stocks |
4-axis scoring
| Axis | Full score | Judgment condition |
|---|---|---|
| RSI strength | 25 | RSI ≧ 60 |
| MACD Cross | 25 | MACD > signal line (bullish cross) |
| Momentum rate | 25 | 20-day price change rate exceeds the threshold |
| Volume surge | 25 | 5-day average volume / 20-day average volume ≧ 1.3 |
A total score ≧ 40 is judged as a momentum stock.
Output
Display the results in Markdown table format. In EquityQuery mode, a sector column is added.
Output columns in EquityQuery mode
Rank / Stock / Sector / Stock price / PER / PBR / Dividend yield / ROE / Score
Output columns in Legacy mode
Rank / Stock / Stock price / PER / PBR / Dividend yield / ROE / Score
Output columns in Pullback mode
Rank / Stock / Stock price / PER / Pullback % / RSI / Volume ratio / SMA50 / SMA200 / Score
Output columns in Alpha mode
Rank / Stock / Stock price / PER / PBR / Undervalued / Change / Overall / Pullback / A / Acceleration / FCF / ROE trend
Output columns in Growth mode
Rank / Stock / Sector / Stock price / PER / PBR / EPS growth / Sales growth / ROE
Output columns in Trending mode
Rank / Stock / Reason for being talked about / Stock price / PER / PBR / Dividend yield / ROE / Score / Judgment
Output columns in Contrarian mode
Rank / Stock / Stock price / PER / PBR / RSI / SMA200 deviation / Tech / Value / Fundamentals / Overall / Judgment
Output columns in Momentum mode
Rank / Stock / Stock price / PER / RSI / MACD / Momentum rate / Volume ratio / Momentum score / Overall score / Classification
Output columns in Shareholder Return mode
Rank / Stock / Stock price / Dividend yield / Share buyback yield / Total payout ratio / Stability / ROE / PER
Execution example
# 日本の割安株（デフォルト）
python3 .../run_screen.py --region japan --preset value
# 米国の高配当テクノロジー株
python3 .../run_screen.py --region us --preset high-dividend --sector Technology
# 香港のバリュー株
python3 .../run_screen.py --region hk --preset value
# ASEAN の成長バリュー株（sg, th, my, id, ph を順次実行）
python3 .../run_screen.py --region asean --preset growth-value
# Legacy モードで米国株をスクリーニング
python3 .../run_screen.py --region us --preset value --mode legacy
# 日本株の押し目買い候補
python3 .../run_screen.py --region japan --preset pullback
# 日本株のアルファシグナル（割安＋変化＋押し目）
python3 .../run_screen.py --region japan --preset alpha
# 米国株のアルファシグナル
python3 .../run_screen.py --region us --preset alpha
# 日本の割安株 + 押し目フィルタ
python3 .../run_screen.py --region japan --preset value --with-pullback
# 米国の高配当株 + 押し目フィルタ
python3 .../run_screen.py --region us --preset high-dividend --with-pullback
# X (Twitter) で話題の日本株
python3 .../run_screen.py --region japan --preset trending
# X で話題のAI関連米国株
python3 .../run_screen.py --region us --preset trending --theme "AI"
# X で話題の半導体関連銘柄
python3 .../run_screen.py --region japan --preset trending --theme "半導体"
# 日本の長期投資候補
python3 .../run_screen.py --region japan --preset long-term
# 米国の長期投資候補
python3 .../run_screen.py --region us --preset long-term
# 日本の純成長株（割安制約なし）
python3 .../run_screen.py --region japan --preset growth
# 米国の純成長株
python3 .../run_screen.py --region us --preset growth
# 日本の高還元株
python3 .../run_screen.py --region japan --preset shareholder-return
# 米国の高還元株
python3 .../run_screen.py --region us --preset shareholder-return
# 米国の高成長株（利益不問・PSR基準）
python3 .../run_screen.py --region us --preset high-growth
# 日本の高成長株
python3 .../run_screen.py --region japan --preset high-growth
# AI関連の割安株（米国）
python3 .../run_screen.py --region us --preset value --theme ai
# 半導体の高成長株（米国）
python3 .../run_screen.py --region us --preset high-growth --theme ai
# 防衛関連株（米国・割安）
python3 .../run_screen.py --region us --preset value --theme defense
# EV関連の成長株（米国）
python3 .../run_screen.py --region us --preset growth --theme ev
# バイオテックの高成長株
python3 .../run_screen.py --region us --preset high-growth --theme biotech
# 日本の小型急成長株（時価総額1000億以下）
python3 .../run_screen.py --region japan --preset small-cap-growth
# 米国の小型急成長株（時価総額$1B以下に自動調整）
python3 .../run_screen.py --region us --preset small-cap-growth
# AI関連の小型成長株
python3 .../run_screen.py --region us --preset small-cap-growth --theme ai
# 日本の逆張り候補（売られすぎ × ファンダ堅調）
python3 .../run_screen.py --region japan --preset contrarian
# 米国の逆張り候補
python3 .../run_screen.py --region us --preset contrarian
# テクノロジーセクターの逆張り候補
python3 .../run_screen.py --region japan --preset contrarian --sector Technology
# 日本の急騰・モメンタム銘柄（安定加速のみ、デフォルト）
python3 .claude/skills/screen-stocks/scripts/run_screen.py --preset momentum --region japan --top 10
# 米国の急騰銘柄を含むモメンタムスクリーニング（急騰・過熱も含む）
python3 .claude/skills/screen-stocks/scripts/run_screen.py --preset momentum --region us --top 5 --submode surge
# テクノロジーセクターのモメンタム銘柄
python3 .../run_screen.py --region japan --preset momentum --sector Technology
Annotation feature (KIK-418/419)
Markers based on investment notes and sale history are automatically added to screening results.
Marker legend
| Marker | Meaning | Trigger |
|---|---|---|
| ⚠️ | Concern note present | Investment note type=concern |
| 📝 | Lesson note present | Investment note type=lesson |
| 👀 | Wait and see | Investment note type=observation + keywords such as "pass" or "wait" |
Automatic exclusion of sold stocks
Stocks sold within the last 90 days are automatically excluded from the screening results. The number of exclusions will be shown in a message.
Data Source
- Neo4j Knowledge Graph (preferred)
- JSON file (fallback when Neo4j is not connected)
Prior Knowledge Integration Rules (KIK-466)
If the output of get_context.py contains the following, integrate it with the screening results and respond:
- Frequent stocks (SURFACED 3 or more times): "Top-ranked 3 times in a row → stably undervalued. Detailed report recommended"
- Held stocks: If screening results include held stocks, "Currently held: material for considering additional investment"
- Concern notes: If there is a concern note for a resulting stock, append "⚠️ Concern note present"
- Sold: If a sold stock appears in the top ranks again, "Previously sold → consider re-entry"