longbridge-technical

name	longbridge-technical
description	Core technical-indicator signal engine for stocks listed in HK / US / A-share / Singapore via Longbridge Securities. Computes and interprets MACD, KDJ, RSI, Bollinger Bands, EMA, ADX, and OBV from OHLCV data; combines multi-dimensional votes (trend / mean-reversion / volume-price) to produce a composite buy / sell / neutral signal. Triggers: "技术指标", "MACD", "KDJ", "RSI", "布林带", "布林线", "EMA", "ADX", "OBV", "金叉", "死叉", "超买", "超卖", "技术分析", "趋势指标", "量价", "技術指標", "布林帶", "技術分析", "超買", "超賣", "technical indicator", "MACD signal", "KDJ overbought", "RSI oversold", "Bollinger Bands", "moving average", "golden cross", "death cross", "technical analysis".
license	MIT
metadata	{"author":"longbridge","version":"1.0.0","risk_level":"read_only","requires_login":false,"default_install":true,"requires_mcp":false,"tier":"read"}

longbridge-technical

Computes seven classic technical indicators from 200 days of OHLCV data and produces a composite buy / sell / neutral signal via a multi-dimensional voting mechanism.

Response language: match the user's input language — Simplified Chinese / Traditional Chinese / English.

When to use

• "NVDA MACD 金叉了吗", "TSLA RSI 超买了吗", "700.HK 技术指标怎么看"
• "600519.SH 布林带收口", "AAPL 技術分析信號", "ADX 趋势强吗"
• "technical analysis for NVDA", "is TSLA overbought on RSI", "MACD signal"

Workflow

1. Resolve the symbol to <CODE>.<MARKET> format.
2. Fetch 200 daily candles:

   longbridge kline <SYMBOL> --period day --format json   # run --help for available flags
   

3. Run the Python analysis below to compute all indicators and their individual votes.
4. Report each indicator's current value and signal, then summarise with the composite vote tally.

CLI

longbridge kline NVDA.US   --period day --format json   # run --help for available flags
longbridge kline 700.HK    --period day --format json
longbridge kline 600519.SH --period day --format json

Run longbridge kline --help to verify current flag names and defaults.

Python analysis

import pandas as pd, json, sys

data = json.loads(sys.stdin.read())
df = pd.DataFrame(data)
df = df.rename(columns={"open":"o","high":"h","low":"l","close":"c","volume":"v"})
df[["o","h","l","c","v"]] = df[["o","h","l","c","v"]].apply(pd.to_numeric)

votes = {}

# EMA helper
def ema(s, n): return s.ewm(span=n, adjust=False).mean()

# --- MACD (12, 26, 9) ---
ema12 = ema(df["c"], 12); ema26 = ema(df["c"], 26)
macd = ema12 - ema26; signal = ema(macd, 9); hist = macd - signal
votes["MACD"] = +1 if hist.iloc[-1] > 0 and hist.iloc[-1] > hist.iloc[-2] else (
                -1 if hist.iloc[-1] < 0 and hist.iloc[-1] < hist.iloc[-2] else 0)

# --- RSI (14) ---
delta = df[].diff(); gain = delta.clip(lower=); loss = (-delta).clip(lower=)
avg_g = gain.ewm(alpha=/, adjust=).mean()
avg_l = loss.ewm(alpha=/, adjust=).mean()
rsi =  -  / ( + avg_g / avg_l.replace(, ))
rsi_last = rsi.iloc[-]
votes[] = -  rsi_last >   (+  rsi_last <   )


low9  = df[].rolling().(); high9 = df[].rolling().()
rsv   = (df[] - low9) / (high9 - low9 + ) * 
K = rsv.ewm(alpha=/, adjust=).mean()
D = K.ewm(alpha=/, adjust=).mean()
J = *K - *D
votes[] = +  J.iloc[-] <   (-  J.iloc[-] >   )


mid = df[].rolling().mean(); std = df[].rolling().std()
upper_bb = mid + *std; lower_bb = mid - *std
c_last = df[].iloc[-]
votes[] = +  c_last < lower_bb.iloc[-]  (
                     -  c_last > upper_bb.iloc[-]  )


e50 = ema(df[], ); e200 = ema(df[], )
votes[] = +  e50.iloc[-] > e200.iloc[-]  -


tr = pd.concat([df[]-df[], (df[]-df[].shift()).(),
                (df[]-df[].shift()).()], axis=).(axis=)
dm_plus  = (df[]-df[].shift()).clip(lower=)
dm_minus = (df[].shift()-df[]).clip(lower=)
atr14 = tr.ewm(alpha=/, adjust=).mean()
di_plus  =  * dm_plus.ewm(alpha=/, adjust=).mean() / atr14
di_minus =  * dm_minus.ewm(alpha=/, adjust=).mean() / atr14
dx = (di_plus - di_minus).() / (di_plus + di_minus + ) * 
adx = dx.ewm(alpha=/, adjust=).mean()

votes[] = +  adx.iloc[-] >   di_plus.iloc[-] > di_minus.iloc[-]  (
               -  adx.iloc[-] >   di_minus.iloc[-] > di_plus.iloc[-]  )


obv = (df[] * df[].diff().apply( x:   x >   (-  x <   ))).cumsum()
votes[] = +  obv.iloc[-] > obv.iloc[-]  (-  obv.iloc[-] < obv.iloc[-]  )


total = (votes.values())
composite =   total >=   (  total <= -  )

()
()
()
()
()
()
()
()

Output

Present a table of indicator values, individual votes (+1 / 0 / -1), and a composite summary row. End with a one-sentence interpretation in the user's language.

指标 / 指標 / Indicator	当前值 / 當前值 / Value	信号 / 訊號 / Signal
MACD	hist 值	多/空/中性
RSI(14)	数值	超卖/中性/超买
KDJ J	数值	超卖/中性/超买
Bollinger	价格位置	超卖/中性/超买
EMA 50/200	多空排列	多头/空头
ADX	趋势强度	趋势/震荡
OBV	5日趋势	流入/流出

Error handling

Situation	简体回复 / 繁體回覆 / English reply
command not found: longbridge	请安装 longbridge-terminal / 請安裝 longbridge-terminal / Install longbridge-terminal first
stderr not logged in / unauthorized	请运行 longbridge auth login / 請執行 longbridge auth login / Run longbridge auth login
Other stderr	直接展示错误信息 / 直接顯示錯誤訊息 / Surface error verbatim

MCP fallback

When the CLI is unavailable, fall back to the MCP server. Discover available tools from the MCP server's tool list at runtime.

Related skills

• longbridge-kline — raw OHLCV data and charting
• longbridge-candlestick — K-line pattern recognition
• longbridge-ichimoku — Ichimoku Cloud system
• longbridge-capital-flow — intraday capital-flow signals