vulcan-technical-analysis

name	vulcan-technical-analysis
version	1.0.0
description	Technical indicators (SMA, EMA, RSI, MACD, BBands, ATR, VWAP, ADX, Stoch) and trigger evaluation over Phoenix candle history.
metadata	{"openclaw":{"category":"finance"},"requires":{"bins":["vulcan"],"skills":["vulcan","vulcan-market-intel"]}}

vulcan-technical-analysis

Use this skill when you need a derived read on price — momentum, trend strength, volatility, or mean-reversion — rather than just raw candles. Backed by the kand TA library; data comes from the same Phoenix candle endpoint as vulcan_market_candles.

When To Use Which Indicator

Question	Indicator	Default period	Primary key
Is price trending up or down?	sma, ema	20	sma / ema
Is momentum overbought or oversold?	rsi	14	rsi
Is momentum accelerating or fading?	macd	(12/26/9)	macd, signal, hist
How volatile is price relative to itself?	bbands	20	upper, middle, lower
What is realized volatility (absolute)?	atr	14	atr
Where is the volume-weighted average?	vwap	n/a	vwap
How strong is the trend (any direction)?	adx	14	adx
Is short-term momentum at an extreme?	stoch	14 (k_slow=3, d=3)	k, d

Three Tools

1. Compute a single indicator

vulcan_ta_compute → { symbol: "SOL", indicator: "rsi", timeframe: "1h", period: 14 }
vulcan_ta_compute → { symbol: "SOL", indicator: "macd", timeframe: "4h", params: { fast: 12, slow: 26, signal: 9 } }
vulcan_ta_compute → { symbol: "SOL", indicator: "bbands", timeframe: "1h", params: { dev_up: 2, dev_down: 2 } }

Returns the full series aligned to the candle window plus a summary.verdict line. Use summary.latest for the latest reading.

2. Evaluate a trigger spec

vulcan_ta_signal → {
  symbol: "SOL",
  spec: { indicator: "rsi", timeframe: "1h", op: "lt", threshold: 30 }
}

Ops: lt, lte, gt, gte, crosses_above, crosses_below. The crosses_* ops require two consecutive non-NaN values on opposite sides of the threshold. Use the optional key field to target a non-primary series (e.g. "key": "hist" for the MACD histogram, "key": "d" for the Stochastic %D line).

3. Bundled report

vulcan_ta_report → { symbol: "SOL", timeframe: "1h" }

One call returns RSI + MACD + BBands + ATR + ADX. Each indicator ships three agent-ready blocks:

• latest — named values from the most recent bar (e.g. bbands.latest.upper, bbands.latest.lower, macd.latest.signal, atr.latest.atr). Use these instead of scanning points.
• signals — derived state: rsi.signals.state (oversold|bearish|neutral|bullish|overbought), bbands.signals.state + position_in_band + width_pct, adx.signals.trend_strength (weak|emerging|strong|very_strong), atr.signals.atr_pct_of_price, macd.signals.momentum + optional recent_cross.
• summary — latest, min, max, mean, and the human-readable verdict line.

Per-bar history is omitted by default. Pass points_limit: 20 (or any positive integer) to opt in to the last N bars; signals and summary are computed from the full untrimmed series regardless. For the full series of a single indicator, use vulcan_ta_compute instead.

Trigger Patterns

Mean-reversion long entry:

spec: { indicator: "rsi", timeframe: "1h", op: "lt", threshold: 30 }

Trend-following exit (momentum fading):

spec: { indicator: "macd", timeframe: "1h", op: "crosses_below", threshold: 0, key: "hist" }

Volatility-expansion pause for a TWAP/grid:

# pause if ATR doubles versus where it was when you started
spec: { indicator: "atr", timeframe: "15m", op: "gt", threshold: <2x launch ATR> }

Strong-trend gate (avoid mean-reversion entries in trending markets):

spec: { indicator: "adx", timeframe: "1h", op: "lt", threshold: 25 }

Notes And Gotchas

• Warmup: each indicator needs N candles before it produces a non-NaN value. The tool fetches max(limit, min_warmup + 5) candles automatically; if the API returns fewer, you'll see INDICATOR_WARMUP_INSUFFICIENT — increase limit or use a shorter period.
• Timeframe selection: 1m/5m are noisy for RSI/MACD; prefer 15m/1h for swing decisions and 4h/1d for trend context.
• NaN handling: leading NaNs are expected. Triggers only fire on valid (non-NaN) latest values.
• For a full TA-driven strategy with declarative rules (open/close/reduce on conditions like EMA cross, RSI mean-reversion, MACD trend follow), use the vulcan-ta-strategy skill — that runner owns the loop, persists state, and handles cooldowns automatically. This skill is for ad-hoc indicator reads and trigger evaluation.