pullback-trading

Name: pullback-trading
Author: SKE-Labs

15stars2forks

How to install pullback-trading?

ske-labs/agent-trading-skills/pullback-trading

$npx -y skills add SKE-Labs/agent-trading-skills --skill pullback-trading

Installs into the current project.

Summary

Enter trends on price retracements to key levels. Use when trading with the trend, finding high R:R entries, or timing entries in established trends. Trend confirmation via EMA slope (no ADX).

Files · 1

View on GitHub

SKILL.md

1	# Pullback Trading
2
3	Enter established trends during temporary retracements for optimal risk/reward.
4
5	## Identification
6
7	### Pullback Levels
8
9	\| Level \| Depth \| Trend Strength \|
10	\| --- \| --- \| --- \|
11	\| 20/50 EMA \| Dynamic \| Strong (shallow) \|
12	\| Normalized retracement band \| Shallow/moderate/deep \| Calibrate by instrument and regime \|
13	\| Previous S/R flip \| Variable \| Structure-based \|
14
15	### Entry Confirmation (require before entering)
16
17	- Reversal candlestick pattern (hammer, engulfing)
18	- Momentum indicator turning in trend direction
19	- Volume decrease during pullback, increase on resumption
20
21	## Workflow
22
23	### 1. Confirm Trend on Higher Timeframe
24
25	`get_indicators` returns one indicator per call — fetch each separately:
26	```
27	get_candles(symbol=<symbol>, exchange=<exchange>, interval=<htf_interval>, count=120)
28	get_indicators(indicator_code="ema_21", symbol=<symbol>, exchange=<exchange>, interval=<htf_interval>, count=120)
29	get_indicators(indicator_code="ema_50", symbol=<symbol>, exchange=<exchange>, interval=<htf_interval>, count=120)
30	```
31
32	Then measure a five-bar EMA50 slope and normalize it by price or ATR. Five bars is a modeling choice, not a canonical horizon:
33	```
34	execute python3 -c "ema50=[<...>]; s=(ema50[-1]-ema50[-5])/ema50[-5]*100; print(f'htf_ema50_slope_pct={s:.3f}')"
35	```
36
37	Define bullish and bearish stack rules plus slope thresholds from training data, then freeze them. Track bars since the trend trigger as a candidate maturity feature; do not label the first pullback highest probability without held-out evidence.
38
39	### 2. Get Price Data and Identify Pullback
40
41	```
42	get_candles(symbol=<symbol>, exchange=<exchange>, interval=<primary_interval>, count=60)
43	```
44
45	Use an objective swing rule for HH/HL or LH/LL structure. Measure retracement as a continuous percentage of the prior impulse and test bands; moving averages, prior swings, and Fibonacci coordinates are candidate reference levels.
46
47	### 3. Check Momentum at Pullback Level
48
49	```
50	get_indicators(indicator_code="rsi_21", symbol=<symbol>, exchange=<exchange>, interval=<primary_interval>, count=20)
51	get_indicators(indicator_code="macd_fast", symbol=<symbol>, exchange=<exchange>, interval=<primary_interval>, count=20)
52	```
53
54	Cite RSI / MACD as a 3-5 value progression (latest value alone is folklore):
55	- Uptrend pullback example: `RSI21 47.8 (62.4 → 55.1 → 49.0 → 47.8)`. Define any RSI band and MACD sign rule before testing; neither 50 nor a Fibonacci ratio is a natural boundary.
56	- Downtrend pullback: mirror — RSI pulling up from low extreme toward 50, MACD histogram shrinking from negative toward zero but NOT crossing into positive.
57
58	### 4. Mark Entry Zone
59
60	```
61	draw_chart_analysis(action="create", drawing={
62	"type": "demand",
63	"points": [
64	{"time": <pullback_start>, "price": <fib_38>},
65	{"time": <pullback_end>, "price": <fib_62>}
66	],
67	"options": {"text": "Pullback Entry Zone (38-62%)"}
68	})
69	```
70
71	### 5. Report to Orchestrator
72
73	Trend direction and strength, pullback depth (which Fib level), confirmation signals, entry level, stop below pullback low, target at previous swing high/low.
74
75	## Evidence and Validation
76
77	- Treat the setup as a testable hypothesis, not a prediction. Define thresholds, entry, invalidation, and exit before evaluating outcomes.
78	- Calibrate on the same instrument, venue, session, and timeframe. Use closed candles and a held-out or walk-forward sample; record every variant tried.
79	- Include spread, fees, slippage, borrow or funding, partial fills, and latency. Reject the setup when net expectancy is not positive or depends on one narrow parameter.
80	- Return observed inputs, missing data, cost assumptions, entry, invalidation, exit, and a valid, watch, or no-trade status.
81	- Research basis: The [technical-analysis evidence review](https://papers.ssrn.com/sol3/Delivery.cfm/SSRN_ID603481_code17745.pdf?abstractid=603481) finds mixed results and recurring data-snooping and transaction-cost problems; calibrate pullback rules rather than treating ratios as natural laws.
82
83	## Key Rules
84
85	- Require the predeclared trend and retracement conditions before evaluating an entry.
86	- Define reversal confirmation numerically and use a closed bar.
87	- Treat deeper retracement as a continuous risk feature, not a law at 61.8%.
88	- Put invalidation beyond the structural pullback extreme plus a calibrated volatility/tick buffer, then resize the position.
89	- Test trend age and pullback sequence out of sample rather than assuming the first is best.
90	- Compute EMA slope via `execute`; do NOT call ADX / DMI / Supertrend (API-billed, not in the free whitelist)
91
92	## Related Skills
93
94	- momentum-trading -- pullbacks occur within momentum moves
95	- breakout-trading -

People who liked this also viewed

6 picks

Type

pullback-trading

15stars2forks

View on GitHub

How to install pullback-trading?

ske-labs/agent-trading-skills/pullback-trading

$npx -y skills add SKE-Labs/agent-trading-skills --skill pullback-trading

Installs into the current project.

Summary

Enter trends on price retracements to key levels. Use when trading with the trend, finding high R:R entries, or timing entries in established trends. Trend confirmation via EMA slope (no ADX).

Files · 1

View on GitHub

SKILL.md

1	# Pullback Trading
2
3	Enter established trends during temporary retracements for optimal risk/reward.
4
5	## Identification
6
7	### Pullback Levels
8
9	\| Level \| Depth \| Trend Strength \|
10	\| --- \| --- \| --- \|
11	\| 20/50 EMA \| Dynamic \| Strong (shallow) \|
12	\| Normalized retracement band \| Shallow/moderate/deep \| Calibrate by instrument and regime \|
13	\| Previous S/R flip \| Variable \| Structure-based \|
14
15	### Entry Confirmation (require before entering)
16
17	- Reversal candlestick pattern (hammer, engulfing)
18	- Momentum indicator turning in trend direction
19	- Volume decrease during pullback, increase on resumption
20
21	## Workflow
22
23	### 1. Confirm Trend on Higher Timeframe
24
25	`get_indicators` returns one indicator per call — fetch each separately:
26	```
27	get_candles(symbol=<symbol>, exchange=<exchange>, interval=<htf_interval>, count=120)
28	get_indicators(indicator_code="ema_21", symbol=<symbol>, exchange=<exchange>, interval=<htf_interval>, count=120)
29	get_indicators(indicator_code="ema_50", symbol=<symbol>, exchange=<exchange>, interval=<htf_interval>, count=120)
30	```
31
32	Then measure a five-bar EMA50 slope and normalize it by price or ATR. Five bars is a modeling choice, not a canonical horizon:
33	```
34	execute python3 -c "ema50=[<...>]; s=(ema50[-1]-ema50[-5])/ema50[-5]*100; print(f'htf_ema50_slope_pct={s:.3f}')"
35	```
36
37	Define bullish and bearish stack rules plus slope thresholds from training data, then freeze them. Track bars since the trend trigger as a candidate maturity feature; do not label the first pullback highest probability without held-out evidence.
38
39	### 2. Get Price Data and Identify Pullback
40
41	```
42	get_candles(symbol=<symbol>, exchange=<exchange>, interval=<primary_interval>, count=60)
43	```
44
45	Use an objective swing rule for HH/HL or LH/LL structure. Measure retracement as a continuous percentage of the prior impulse and test bands; moving averages, prior swings, and Fibonacci coordinates are candidate reference levels.
46
47	### 3. Check Momentum at Pullback Level
48
49	```
50	get_indicators(indicator_code="rsi_21", symbol=<symbol>, exchange=<exchange>, interval=<primary_interval>, count=20)
51	get_indicators(indicator_code="macd_fast", symbol=<symbol>, exchange=<exchange>, interval=<primary_interval>, count=20)
52	```
53
54	Cite RSI / MACD as a 3-5 value progression (latest value alone is folklore):
55	- Uptrend pullback example: `RSI21 47.8 (62.4 → 55.1 → 49.0 → 47.8)`. Define any RSI band and MACD sign rule before testing; neither 50 nor a Fibonacci ratio is a natural boundary.
56	- Downtrend pullback: mirror — RSI pulling up from low extreme toward 50, MACD histogram shrinking from negative toward zero but NOT crossing into positive.
57
58	### 4. Mark Entry Zone
59
60	```
61	draw_chart_analysis(action="create", drawing={
62	"type": "demand",
63	"points": [
64	{"time": <pullback_start>, "price": <fib_38>},
65	{"time": <pullback_end>, "price": <fib_62>}
66	],
67	"options": {"text": "Pullback Entry Zone (38-62%)"}
68	})
69	```
70
71	### 5. Report to Orchestrator
72
73	Trend direction and strength, pullback depth (which Fib level), confirmation signals, entry level, stop below pullback low, target at previous swing high/low.
74
75	## Evidence and Validation
76
77	- Treat the setup as a testable hypothesis, not a prediction. Define thresholds, entry, invalidation, and exit before evaluating outcomes.
78	- Calibrate on the same instrument, venue, session, and timeframe. Use closed candles and a held-out or walk-forward sample; record every variant tried.
79	- Include spread, fees, slippage, borrow or funding, partial fills, and latency. Reject the setup when net expectancy is not positive or depends on one narrow parameter.
80	- Return observed inputs, missing data, cost assumptions, entry, invalidation, exit, and a valid, watch, or no-trade status.
81	- Research basis: The [technical-analysis evidence review](https://papers.ssrn.com/sol3/Delivery.cfm/SSRN_ID603481_code17745.pdf?abstractid=603481) finds mixed results and recurring data-snooping and transaction-cost problems; calibrate pullback rules rather than treating ratios as natural laws.
82
83	## Key Rules
84
85	- Require the predeclared trend and retracement conditions before evaluating an entry.
86	- Define reversal confirmation numerically and use a closed bar.
87	- Treat deeper retracement as a continuous risk feature, not a law at 61.8%.
88	- Put invalidation beyond the structural pullback extreme plus a calibrated volatility/tick buffer, then resize the position.
89	- Test trend age and pullback sequence out of sample rather than assuming the first is best.
90	- Compute EMA slope via `execute`; do NOT call ADX / DMI / Supertrend (API-billed, not in the free whitelist)
91
92	## Related Skills
93
94	- momentum-trading -- pullbacks occur within momentum moves
95	- breakout-trading -

People who liked this also viewed

6 picks

Type