opencode-skills-collection 4.0.69 → 4.0.70

package/bundled-skills/longbridge-quant/references/hedging.md

@@ -0,0 +1,136 @@
+# longbridge-hedging
+
+Design and evaluate hedging strategies for a portfolio or single position using Longbridge market data — from simple Beta hedges to options-based protection and cross-asset tail-risk hedges.
+
+## Workflow
+
+### Step 1 — Identify hedge objective
+
+Clarify with the user:
+
+- What is being hedged: single position, portfolio, or sector exposure?
+- Risk to hedge: market Beta, tail event, currency, or volatility?
+- Hedge horizon: days, weeks, or months?
+- Cost tolerance: zero-cost (collar) or willing to pay premium?
+
+### Step 2 — Fetch data
+
+```bash
+longbridge kline --help
+longbridge option --help
+
+# Beta calculation (60-day daily returns)
+longbridge kline <SYMBOL> --period day --count 60 --format json
+
+# Option chain for hedge instruments
+longbridge option chain <SYMBOL> --format json
+
+# Current portfolio positions (requires login with trade scope)
+longbridge positions --format json
+```
+
+### Step 3 — Beta hedge
+
+**Portfolio Beta**:
+
+```
+β_portfolio = Σ(w_i × β_i)
+```
+
+Compute individual Beta for each holding from 60-day returns vs benchmark (SPX / HSI / CSI300). Fetch benchmark kline with `longbridge kline <BENCHMARK> --period day --count 60 --format json`.
+
+**Hedge ratio (index futures or inverse ETF)**:
+
+```
+Contracts needed = (Portfolio Value × β_portfolio) / (Futures Price × Contract Multiplier)
+```
+
+Present: number of contracts, hedge cost, and residual Beta after hedge.
+
+### Step 4 — Options-based protection
+
+**Protective Put** (保护性看跌期权介绍):
+
+- 原理：持有正股的同时持有看跌期权；当标的价格下跌时，期权价值上升，可对冲下行风险。常见做法是选择平值（ATM）或略虚值（OTM）的看跌期权。
+- Cost = put premium; protection kicks in below strike.
+- Effective floor = Strike − Premium paid.
+- 具体期权合约是否适用，请根据自身持仓情况和风险偏好独立判断。
+- Fetch available strikes: `longbridge option chain <SYMBOL> --format json`.
+
+**Collar Strategy** (zero-cost or near-zero):
+
+- Buy OTM put (downside protection) + sell OTM call (cap upside).
+- Net premium ≈ 0 if call premium offsets put premium.
+- Present: put strike, call strike, net cost, max gain, max loss.
+
+**Selection criteria**:
+| Criterion | Protective Put | Collar |
+|---|---|---|
+| Upside retention | Full | Capped at call strike |
+| Cost | Premium paid | Near zero |
+| Best for | Bullish with hedge need | Neutral/mild bearish |
+
+### Step 5 — Tail risk hedges
+
+| Tool                     | Instrument            | Mechanism                          |
+| ------------------------ | --------------------- | ---------------------------------- |
+| VIX calls                | UVXY.US / VIX options | Profit from volatility spike       |
+| Gold                     | GLD.US / 518880.SH    | Safe-haven in risk-off             |
+| Long-dated US Treasuries | TLT.US                | Negative correlation with equities |
+| Put on index             | SPY puts / HSI puts   | Direct market hedge                |
+
+Note: fetch current price and recent kline for any hedge instrument before recommending.
+
+### Step 6 — Currency hedge
+
+For HK/US cross-currency portfolios:
+
+- USD/HKD is pegged — minimal FX risk.
+- CNY exposure: use offshore RMB (CNH) forwards or futures.
+- Non-HKD Asian exposure: fetch FX rate via `longbridge fx --format json` (verify flag with `--help`).
+
+Present notional hedge amount, instrument, tenor, and estimated cost.
+
+### Step 7 — Hedge cost assessment
+
+```
+Cost efficiency = Protection value / Premium paid
+```
+
+Present: premium as % of protected notional, breakeven move, and expected cost per 1% of downside protection.
+
+## CLI
+
+```bash
+longbridge kline --help
+longbridge option --help
+longbridge positions --help
+
+longbridge kline <SYMBOL> --period day --count 60 --format json
+longbridge option chain <SYMBOL> --format json
+longbridge positions --format json
+```
+
+## Output
+
+Present:
+
+1. Hedge objective summary.
+2. Recommended strategy with rationale.
+3. Implementation details (strikes, contracts, premium).
+4. Cost vs protection table.
+5. Scenarios: portfolio value if market falls 10% / 20% with and without hedge.
+6. Caveats (basis risk, early exercise for American options, liquidity).
+
+Always note: hedging reduces risk but also limits upside.
+
+> 以上内容仅供参考，不构成投资建议。投资决策请结合自身风险承受能力独立判断。/ The above is for reference only and does not constitute investment advice. Please make investment decisions independently based on your own risk tolerance.
+
+## Error handling
+
+| Situation                       | 简体回复                                                  | 繁體回覆                                                  | English reply                                                      |
+| ------------------------------- | --------------------------------------------------------- | --------------------------------------------------------- | ------------------------------------------------------------------ |
+| `command not found: longbridge` | 请安装 longbridge-terminal 或检查 MCP 配置。              | 請安裝 longbridge-terminal 或檢查 MCP 配置。              | Install longbridge-terminal or check MCP config.                   |
+| stderr: `not logged in`         | 请运行 `longbridge auth login`（需 Trade 权限查看持仓）。 | 請執行 `longbridge auth login`（需 Trade 權限查看持倉）。 | Run `longbridge auth login` (Trade scope needed for positions).    |
+| No option chain data            | 该标的无期权数据，请尝试对应指数期权或 ETF 期权。         | 該標的無期權數據，請嘗試指數或 ETF 期權。                 | No option chain for this symbol; try index or ETF options instead. |
+| Negative or missing Beta        | Beta 数据不足，将使用市值加权 Beta=1 作为默认值。         | Beta 數據不足，使用 Beta=1 作為默認值。                   | Insufficient Beta data; defaulting to Beta = 1.                    |

package/bundled-skills/longbridge-quant/references/ml-strategy.md

@@ -0,0 +1,77 @@
+# longbridge-ml-strategy
+
+Walk-forward machine-learning framework for stock direction prediction. Fetches historical OHLCV data, engineers technical features, trains a rolling classifier (Random Forest or Gradient Boosting), generates probabilistic buy/sell signals, and evaluates backtest performance.
+
+## Dependencies
+
+Requires: `scikit-learn`, `pandas`, `numpy` (usually pre-installed).
+Optional: `xgboost` or `lightgbm` for gradient-boosting models.
+If unavailable, fall back to a simpler logistic-regression model.
+
+## Workflow
+
+1. Fetch 504 daily candles (≈ 2 years):
+   `longbridge kline <SYMBOL> --period day --count 504 --format json`
+
+2. **Feature engineering** (compute on rolling windows):
+   - MACD line and signal (EMA12 − EMA26, signal EMA9)
+   - RSI-14
+   - Bollinger Band width: (upper − lower) / mid, window 20
+   - Volume change rate: (vol*t − vol*{t-5}) / vol\_{t-5}
+   - 5-day price momentum: (close*t / close*{t-5}) − 1
+   - Label: 1 if close\_{t+5} > close_t × 1.01, 0 if < close_t × 0.99, else drop
+
+3. **Walk-forward training**:
+   - Training window: 252 days; retrain every 60 days
+   - Model: `RandomForestClassifier(n_estimators=100)` or `GradientBoostingClassifier`
+   - Predict probability for the current bar
+
+4. **Signal generation**:
+   - prob > 0.60 → 模型上涨概率偏高 / Model upside probability elevated
+   - prob < 0.40 → 模型下跌概率偏高 / Model downside probability elevated
+   - Otherwise → 模型无方向性预测 / No directional signal from model
+
+5. **Backtest metrics** (on out-of-sample predictions):
+   - Win rate (% correct directional calls)
+   - Profit factor (gross profit / gross loss)
+   - Annualised Sharpe ratio (assuming daily rebalance)
+   - Max drawdown
+
+6. **Feature importance**: rank top-5 features by mean decrease in impurity.
+
+Run `longbridge kline --help` to confirm flag names before calling.
+
+## CLI
+
+```bash
+longbridge kline --help
+
+longbridge kline <SYMBOL> --period day --count 504 --format json
+```
+
+## Output
+
+| Metric             | 简体     | 繁體     | English            |
+| ------------------ | -------- | -------- | ------------------ |
+| Current signal     | 当前信号 | 當前訊號 | Current signal     |
+| Signal probability | 预测概率 | 預測概率 | Signal probability |
+| Win rate           | 胜率     | 勝率     | Win rate           |
+| Profit factor      | 盈亏比   | 盈虧比   | Profit factor      |
+| Sharpe ratio       | 夏普比率 | 夏普比率 | Sharpe ratio       |
+| Max drawdown       | 最大回撤 | 最大回撤 | Max drawdown       |
+| Top features       | 重要特征 | 重要特徵 | Top features       |
+
+Output: current signal box → backtest summary table → feature importance list → caveats (past performance, data snooping). Cite **Longbridge Securities** / **数据来源：长桥证券** / **數據來源：長橋證券**.
+
+> 以上内容仅供参考，不构成投资建议。投资决策请结合自身风险承受能力独立判断。
+> The above is for reference only and does not constitute investment advice. Investment decisions should be made based on your own risk tolerance.
+
+## Error handling
+
+| Situation                        | 简体回复                                                         | 繁體回復                                  | English reply                                   |
+| -------------------------------- | ---------------------------------------------------------------- | ----------------------------------------- | ----------------------------------------------- |
+| `command not found: longbridge`  | 回退到 MCP 或提示安装 longbridge-terminal                        | 回退到 MCP 或提示安裝 longbridge-terminal | Fall back to MCP or install longbridge-terminal |
+| `not logged in` / `unauthorized` | 请运行 `longbridge auth login`                                   | 請執行 `longbridge auth login`            | Run `longbridge auth login`                     |
+| `scikit-learn` not found         | 提示 `pip install scikit-learn pandas numpy`，并改用逻辑回归降级 | 提示安裝，降級至邏輯回歸                  | Prompt install; degrade to logistic regression  |
+| Fewer than 252 candles           | 数据不足，无法完成 walk-forward 训练                             | 數據不足                                  | Insufficient data for walk-forward              |
+| Other stderr                     | 直接显示原始错误                                                 | 直接顯示原始錯誤                          | Surface verbatim                                |

package/bundled-skills/longbridge-quant/references/multifactor.md

@@ -0,0 +1,68 @@
+# longbridge-multifactor
+
+Cross-sectional multi-factor quantitative stock selection. Scores a universe of stocks on value, momentum, quality, and low-volatility factors; composites the scores; ranks stocks; and outputs a TopN buy list and bottom-N short list with factor-level attribution.
+
+## Workflow
+
+1. **Get universe**: fetch index constituents:
+   `longbridge constituent <INDEX> --format json`
+   Extract the `stocks` key. If the user provides a custom list, skip this step.
+
+2. **Fetch valuation factors** for each symbol (batched, up to 20 stocks for manageable output):
+   `longbridge calc-index <SYMBOL> --format json`
+   Extract PE, PB, ROE. Value factors: `f_value = 0.5 × (1/PE) + 0.5 × (1/PB)` (normalised).
+
+3. **Fetch price history** for momentum and low-vol:
+   `longbridge kline <SYMBOL> --period day --count 60 --format json`
+   - Momentum: (close_today / close_60d_ago) − 1
+   - Low-volatility: annualised std of last 60 daily returns × √252 (negate: lower HV → higher score)
+
+4. **Standardise** each factor across the universe to Z-scores (subtract mean, divide by std).
+
+5. **Composite score**:
+   - Equal-weight: `score = 0.25 × Z_value + 0.25 × Z_momentum + 0.25 × Z_quality + 0.25 × Z_lowvol`
+   - IC-weighted (if the user specifies): weight each factor by its historical IC (information coefficient); if IC data unavailable, default to equal-weight.
+
+6. **Rank and output**:
+   - Top 20%: buy / long signal
+   - Bottom 20%: avoid / short signal
+   - Display top-10 and bottom-10 stocks with individual factor Z-scores and composite score.
+
+Run `longbridge constituent --help`, `longbridge calc-index --help`, and `longbridge kline --help` to verify current flag names.
+
+## CLI
+
+```bash
+longbridge constituent --help
+longbridge calc-index --help
+longbridge kline --help
+
+longbridge constituent <INDEX>  --format json
+longbridge calc-index <SYMBOL>  --format json
+longbridge kline <SYMBOL> --period day --count 60 --format json
+```
+
+Supported index examples: `HSI.HK`, `SPX.US`, `IXIC.US`, `DJI.US`, `000300.SH`.
+
+## Output
+
+| Column          | 简体            | 繁體            | English          |
+| --------------- | --------------- | --------------- | ---------------- |
+| Composite score | 综合得分        | 綜合得分        | Composite score  |
+| Value Z         | 价值因子 Z 值   | 價值因子 Z 值   | Value Z-score    |
+| Momentum Z      | 动量因子 Z 值   | 動量因子 Z 值   | Momentum Z-score |
+| Quality Z       | 质量因子 Z 值   | 質量因子 Z 值   | Quality Z-score  |
+| Low-vol Z       | 低波动因子 Z 值 | 低波動因子 Z 值 | Low-vol Z-score  |
+| Signal          | 信号            | 訊號            | Signal           |
+
+Output: top-10 / bottom-10 ranked table → factor dispersion summary → composite methodology note. Add caveat that the universe is limited by API throughput. Cite **Longbridge Securities** / **数据来源：长桥证券** / **數據來源：長橋證券**.
+
+## Error handling
+
+| Situation                        | 简体回复                                  | 繁體回復                                  | English reply                                   |
+| -------------------------------- | ----------------------------------------- | ----------------------------------------- | ----------------------------------------------- |
+| `command not found: longbridge`  | 回退到 MCP 或提示安装 longbridge-terminal | 回退到 MCP 或提示安裝 longbridge-terminal | Fall back to MCP or install longbridge-terminal |
+| `not logged in` / `unauthorized` | 请运行 `longbridge auth login`            | 請執行 `longbridge auth login`            | Run `longbridge auth login`                     |
+| `calc-index` returns null PE/PB  | 跳过该标的，标注"数据缺失"                | 跳過該標的，標注"數據缺失"                | Skip symbol; mark as "data missing"             |
+| Universe > 50 stocks             | 自动截取前50只成交额最大的标的            | 自動截取前50只                            | Auto-limit to top-50 by turnover                |
+| Other stderr                     | 直接显示原始错误                          | 直接顯示原始錯誤                          | Surface verbatim                                |

package/bundled-skills/longbridge-quant/references/pairs-trading.md

@@ -0,0 +1,61 @@
+# longbridge-pairs-trading
+
+Statistical-arbitrage strategy for a pair of correlated securities. Tests for cointegration, estimates hedge ratio, computes spread Z-score, and outputs actionable long/short signals with half-life and position sizing guidance.
+
+## Workflow
+
+1. Fetch 252 daily candles for each symbol:
+   ```
+   longbridge kline <SYMBOL_A> --period day --count 252 --format json
+   longbridge kline <SYMBOL_B> --period day --count 252 --format json
+   ```
+2. Align on `time`, drop unmatched rows (different trading calendars).
+3. **Cointegration test (Engle-Granger)**:
+   - OLS regress `ln(close_A)` on `ln(close_B)` → hedge ratio β
+   - Compute residuals (spread) = `ln(close_A) − β × ln(close_B)`
+   - Run ADF test on residuals; if p-value < 0.05, declare cointegrated
+4. **Spread statistics**:
+   - Spread mean μ, std σ
+   - Z-score = (spread_current − μ) / σ
+   - Half-life λ = −ln(2) / OLS*slope of Δspread ~ spread*{t-1} (AR(1))
+5. **Signal**:
+   - Z > 2.0: 价差处于历史高位区间（统计上偏离均值偏大）/ Spread at historical high (statistically elevated)
+   - Z < −2.0: 价差处于历史低位区间（统计上偏离均值偏小）/ Spread at historical low (statistically depressed)
+   - |Z| < 0.5: 价差回归均值区间 / Spread near historical mean
+6. Position sizing: suggest equal-dollar or volatility-scaled sizing; note that execution must be simultaneous.
+
+Run `longbridge kline --help` to confirm current flag names before calling.
+
+## CLI
+
+```bash
+longbridge kline --help
+
+longbridge kline <SYMBOL_A> --period day --count 252 --format json
+longbridge kline <SYMBOL_B> --period day --count 252 --format json
+```
+
+## Output
+
+| Metric                | 简体         | 繁體         | English               |
+| --------------------- | ------------ | ------------ | --------------------- |
+| Hedge ratio β         | 对冲比率     | 對沖比率     | Hedge ratio           |
+| Cointegration p-value | 协整 p 值    | 協整 p 值    | Cointegration p-value |
+| Spread Z-score        | 价差 Z 分    | 價差 Z 分    | Spread Z-score        |
+| Half-life             | 半衰期（天） | 半衰期（天） | Half-life (days)      |
+| Signal                | 交易信号     | 交易訊號     | Trade signal          |
+
+Output: cointegration verdict → spread statistics table → current signal → position guidance. Add a risk note if p-value > 0.05 (not cointegrated). Cite **Longbridge Securities** / **数据来源：长桥证券** / **數據來源：長橋證券**.
+
+> 以上内容仅供参考，不构成投资建议。投资决策请结合自身风险承受能力独立判断。
+> The above is for reference only and does not constitute investment advice. Investment decisions should be made based on your own risk tolerance.
+
+## Error handling
+
+| Situation                        | 简体回复                                  | 繁體回復                                  | English reply                                   |
+| -------------------------------- | ----------------------------------------- | ----------------------------------------- | ----------------------------------------------- |
+| `command not found: longbridge`  | 回退到 MCP 或提示安装 longbridge-terminal | 回退到 MCP 或提示安裝 longbridge-terminal | Fall back to MCP or install longbridge-terminal |
+| `not logged in` / `unauthorized` | 请运行 `longbridge auth login`            | 請執行 `longbridge auth login`            | Run `longbridge auth login`                     |
+| ADF p-value > 0.05               | 两标的未通过协整检验，配对交易风险较高    | 兩標的未通過協整檢驗，配對交易風險較高    | Not cointegrated; pairs trade is high-risk      |
+| Insufficient overlapping dates   | 两标的历史数据重叠不足，无法建立配对      | 兩標的歷史數據重疊不足                    | Insufficient overlapping history                |
+| Other stderr                     | 直接显示原始错误                          | 直接顯示原始錯誤                          | Surface verbatim                                |

package/bundled-skills/longbridge-quant/references/quant-cli.md

@@ -0,0 +1,133 @@
+# quant
+
+Quantitative analysis: run indicator scripts against K-line data
+
+Subcommands: run Example: longbridge quant run TSLA.US --start 2024-01-01 --end 2024-12-31 --script "..." Example: cat script.nv | longbridge quant run TSLA.US --start 2024-01-01 --end 2024-12-31
+
+Usage: longbridge quant [OPTIONS] <COMMAND>
+
+Commands:
+  run   Run a quant indicator script against historical K-line data on the server
+  help  Print this message or the help of the given subcommand(s)
+
+Options:
+      --format <FORMAT>
+          Output format: 'pretty' for human-readable, 'json' for AI agents and scripting
+          
+          [default: pretty]
+          [possible values: table, json]
+
+  -v, --verbose
+          Print verbose request info (host, elapsed) to stderr, prefixed with `*` like curl -v
+
+      --lang <LANG>
+          Language for content fetched from longbridge.com: zh-CN or en. Defaults to system LANG env var, then en
+
+      --schema
+          Show response fields for this command and exit
+
+  -h, --help
+          Print help (see a summary with '-h')
+
+---
+
+## `quant run` — run an indicator script
+
+```
+Run a quant indicator script against historical K-line data on the server
+
+Executes the script server-side and returns the computed indicator/plot values as JSON. Scripts are written in Navi (.nv); pass --language pine for PineScript compatibility.
+
+Periods: 1m  5m  15m  30m  1h  day  week  month  year
+
+Script source (--script takes priority over stdin): --script TEXT   inline script text stdin           cat script.nv | longbridge quant run TSLA.US ...
+
+The optional --input flag accepts a JSON array matching the order of input.*() calls in the script, e.g. --input '[14,2.0]'
+
+Example: longbridge quant run TSLA.US --start 2024-01-01 --end 2024-12-31 --script "..." Example: cat script.nv | longbridge quant run TSLA.US --start 2024-01-01 --end 2024-12-31 Example: longbridge quant run 700.HK --period 1h --start 2024-01-01 --end 2024-06-30 --script "..." --input '[14]' Example: longbridge quant run TSLA.US --start 2024-01-01 --end 2024-12-31 --script "..." --format json Example: longbridge quant run 700.HK --period 1m --start "2024-01-02 09:30" --end "2024-01-02 16:00" --script "..." Example: cat script.pine | longbridge quant run TSLA.US --start 2024-01-01 --end 2024-12-31 --language pine
+
+Usage: longbridge quant run [OPTIONS] --start <START> --end <END> <SYMBOL>
+
+Arguments:
+  <SYMBOL>
+          Symbol in <CODE>.<MARKET> format, e.g. TSLA.US 700.HK
+
+Options:
+      --period <PERIOD>
+          K-line period: 1m 5m 15m 30m 1h day week month year (default: day)
+          
+          [default: day]
+
+      --start <START>
+          Start date/time for the K-line range (local YYYY-MM-DD, local "YYYY-MM-DD HH:MM", or RFC 3339)
+
+      --end <END>
+          End date/time for the K-line range (local YYYY-MM-DD, local "YYYY-MM-DD HH:MM", or RFC 3339)
+
+      --script <SCRIPT>
+          Script text. Omit to read from stdin (e.g. echo "..." | longbridge quant run ...)
+
+      --input <INPUT>
+          Script input values as a JSON array, e.g. '[14,2.0]' Must match the order of input.*() calls in the script
+
+      --language <LANGUAGE>
+          Script language: `navi` (default), or `pine` for PineScript compatibility
+          
+          [default: navi]
+
+      --format <FORMAT>
+          Output format: 'pretty' for human-readable, 'json' for AI agents and scripting
+          
+          [default: pretty]
+          [possible values: table, json]
+
+  -v, --verbose
+          Print verbose request info (host, elapsed) to stderr, prefixed with `*` like curl -v
+
+      --lang <LANG>
+          Language for content fetched from longbridge.com: zh-CN or en. Defaults to system LANG env var, then en
+
+      --schema
+          Show response fields for this command and exit
+
+  -h, --help
+          Print help (see a summary with '-h')
+```
+
+## Example script (Navi)
+
+```nv
+indicator("RSI");
+
+let length = input.int(14, "Length", minval: 1);
+let rsi = ta.rsi(close, length);
+
+plot(rsi, "RSI");
+plot(70.0, "OB");
+plot(30.0, "OS");
+```
+
+```bash
+cat rsi.nv | longbridge quant run AAPL.US --start 2025-01-01 --end 2026-01-31
+```
+
+## Usage patterns
+
+```bash
+# Run an inline script against a symbol
+longbridge quant run TSLA.US --start 2024-01-01 --end 2024-12-31 --script "..."
+
+# Pipe a script file
+cat my_strategy.nv | longbridge quant run NVDA.US --start 2024-01-01 --end 2024-12-31
+
+# Always add --format json for AI agent processing
+longbridge quant run TSLA.US --start 2024-01-01 --end 2024-12-31 --script "..." --format json
+```
+
+## Notes
+
+- Script language: **Navi** (`.nv`) by default; `--language pine` for PineScript compatibility. Any unrecognised value falls back to Navi.
+- Navi syntax and standard library: <https://navi-lang.org> is authoritative. Install its CLI (<https://navi-lang.org/docs/install.md>) and run `navi lint script.nv` before sending — the API only reports script errors as an opaque error code.
+- `--format json` sets `exclude_chart`, so series values are absent from the JSON response; backtest metrics still come back via `.report_json`.
+- Run `longbridge quant run --help` for all current flags
+- Use `longbridge kline` to preview the underlying OHLCV data first

package/bundled-skills/longbridge-quant/references/quant-stats.md

@@ -0,0 +1,150 @@
+# longbridge-quant-stats
+
+Apply rigorous statistical methods to financial time-series data retrieved from Longbridge — test assumptions before modelling, diagnose residuals, and produce statistically sound inferences.
+
+## Prerequisites
+
+```bash
+pip install statsmodels scipy numpy pandas
+```
+
+## Workflow and test catalogue
+
+### Step 1 — Fetch price data
+
+```bash
+longbridge kline --help
+longbridge kline <SYMBOL> --period day --count 252 --format json
+```
+
+Extract the `close` price series. Compute log returns: `r_t = ln(P_t / P_{t-1})`.
+
+### Step 2 — Stationarity: ADF Unit Root Test
+
+**When to use**: Before regression or time-series modelling — most models require stationary series.
+
+**Python (statsmodels)**:
+
+```python
+from statsmodels.tsa.stattools import adfuller
+result = adfuller(series, autolag='AIC')
+# result: (adf_stat, p_value, lags, n_obs, critical_values, icbest)
+```
+
+**Interpretation**:
+
+- p < 0.05 → reject unit root → series is stationary.
+- p ≥ 0.05 → fail to reject → series has unit root → difference the series.
+- Log prices: usually non-stationary. Log returns: usually stationary.
+
+### Step 3 — Cointegration Test
+
+**When to use**: Two non-stationary series may share a long-run equilibrium (pairs trading).
+
+**Engle-Granger (two-series)**:
+
+```python
+from statsmodels.tsa.stattools import coint
+t_stat, p_value, critical_values = coint(series_A, series_B)
+# p < 0.05 → cointegrated
+```
+
+**Johansen (multivariate)**:
+
+```python
+from statsmodels.tsa.vector_ar.vecm import coint_johansen
+result = coint_johansen(df, det_order=0, k_ar_diff=1)
+# trace statistic vs critical values at 90%/95%/99%
+```
+
+Report: test statistic, p-value, critical values, and cointegrating vector.
+
+### Step 4 — GARCH Volatility Modelling
+
+**When to use**: Financial returns show volatility clustering (ARCH effects).
+
+```python
+from arch import arch_model
+model = arch_model(returns * 100, vol='Garch', p=1, q=1)
+res = model.fit(disp='off')
+print(res.summary())
+```
+
+Note: `pip install arch` required in addition to statsmodels.
+
+**Output**: omega, alpha (ARCH), beta (GARCH) coefficients. Persistence = alpha + beta. If > 0.95, volatility is highly persistent.
+
+**ARCH-LM test first** (to verify ARCH effects exist):
+
+```python
+from statsmodels.stats.diagnostic import het_arch
+lm_stat, p_value, f_stat, f_p = het_arch(residuals)
+```
+
+### Step 5 — Regression Diagnostics
+
+After running OLS (`statsmodels.api.OLS`), check:
+
+| Test          | Purpose                             | Command                                                          |
+| ------------- | ----------------------------------- | ---------------------------------------------------------------- |
+| Durbin-Watson | Serial autocorrelation in residuals | `statsmodels.stats.stattools.durbin_watson(resid)`               |
+| Breusch-Pagan | Heteroskedasticity                  | `statsmodels.stats.diagnostic.het_breuschpagan(resid, exog)`     |
+| Jarque-Bera   | Normality of residuals              | `statsmodels.stats.stattools.jarque_bera(resid)`                 |
+| VIF           | Multicollinearity                   | `statsmodels.stats.outliers_influence.variance_inflation_factor` |
+
+Interpret Durbin-Watson: ~2.0 = no autocorrelation; < 1.5 = positive autocorrelation; > 2.5 = negative autocorrelation.
+
+### Step 6 — Bootstrap Confidence Intervals
+
+**When to use**: Non-normal distributions; small samples; estimating CI for Sharpe ratio, IC, or any statistic.
+
+```python
+import numpy as np
+
+def bootstrap_ci(data, stat_fn, n_boot=10000, ci=0.95):
+    boots = [stat_fn(np.random.choice(data, len(data), replace=True))
+             for _ in range(n_boot)]
+    lo = np.percentile(boots, (1 - ci) / 2 * 100)
+    hi = np.percentile(boots, (1 + ci) / 2 * 100)
+    return lo, hi
+
+# Example: Sharpe ratio CI
+sharpe_lo, sharpe_hi = bootstrap_ci(returns, lambda x: x.mean() / x.std() * np.sqrt(252))
+```
+
+### Step 7 — Hypothesis Tests
+
+| Test                | Use case                                    | Function                                     |
+| ------------------- | ------------------------------------------- | -------------------------------------------- |
+| t-test (one sample) | Is mean IC > 0?                             | `scipy.stats.ttest_1samp(ic_series, 0)`      |
+| t-test (two sample) | Is long portfolio return > short portfolio? | `scipy.stats.ttest_ind(long_ret, short_ret)` |
+| F-test / ANOVA      | Are returns different across deciles?       | `scipy.stats.f_oneway(*decile_returns)`      |
+| Mann-Whitney U      | Non-parametric alternative to t-test        | `scipy.stats.mannwhitneyu(a, b)`             |
+
+Always report: test statistic, p-value, degrees of freedom, and conclusion at 5% significance level.
+
+## CLI
+
+```bash
+longbridge kline --help
+longbridge kline <SYMBOL> --period day --count 252 --format json
+```
+
+## Output
+
+For each test present:
+
+1. Test name and null hypothesis.
+2. Test statistic and p-value.
+3. Critical values (where applicable).
+4. Conclusion at 5% significance.
+5. Practical implication for the user's use case.
+
+## Error handling
+
+| Situation                             | 简体回复                                              | 繁體回覆                                              | English reply                                                  |
+| ------------------------------------- | ----------------------------------------------------- | ----------------------------------------------------- | -------------------------------------------------------------- |
+| `command not found: longbridge`       | 请安装 longbridge-terminal 或检查 MCP 配置。          | 請安裝 longbridge-terminal 或檢查 MCP 配置。          | Install longbridge-terminal or check MCP config.               |
+| `ModuleNotFoundError: statsmodels`    | 请运行 `pip install statsmodels scipy numpy pandas`。 | 請執行 `pip install statsmodels scipy numpy pandas`。 | Run `pip install statsmodels scipy numpy pandas`.              |
+| Insufficient data (< 30 observations) | 样本量过小，统计结论可靠性有限，建议延长数据期。      | 樣本量過小，建議延長數據期。                          | Sample too small; extend the data period for reliable results. |
+| ARCH module missing for GARCH         | 请运行 `pip install arch` 以使用 GARCH 模型。         | 請執行 `pip install arch` 以使用 GARCH 模型。         | Run `pip install arch` for GARCH modelling.                    |