Available Skills¶

Complete reference for all Claude Code skills in the RQF-ML system.

PineScript Development¶

/pine-generate¶

Generate production-ready PineScript v6 indicators from specifications or descriptions.

/pine-generate "Detect bearish OB after MSS with FVG confluence"
/pine-generate specs/new_indicator.md

Features:

• Non-repaint compliance built-in
• Proper state variable declarations
• JSON webhook alerts
• Info table display

/pine-analyze¶

Deep analysis of PineScript code for quality, compliance, and bugs.

/pine-analyze strategies/pine_v6/STRAT_02.pine

Checks:

• Repaint detection (critical)
• State machine validation
• Memory/performance issues
• Edge case handling
• Alert completeness
• Code quality

Output: Generates analysis report at strategies/pine_v6/analysis/

/pine-debug¶

Fix PineScript compilation and runtime errors.

/pine-debug STRAT_02.pine "Undeclared identifier 'bullObBox'"
/pine-debug STRAT_02.pine "line 450"

Common Fixes:

• Undeclared identifiers
• Type mismatches
• Array/drawing limits
• request.security scope issues
• Na comparison problems
• Plot/plotarrow conditional errors

/pine-extract¶

Reverse-engineer external PineScript indicators to understand their algorithms.

/pine-extract some_indicator.pine "order block detection"
/pine-extract some_indicator.pine "how does it detect sweeps"

Use For:

• Learning from other indicators
• Understanding unfamiliar code
• Extracting reusable patterns

/pine-patterns¶

Reference library of reusable PineScript v6 patterns.

/pine-patterns pivot
/pine-patterns bos
/pine-patterns state-rqf
/pine-patterns ob
/pine-patterns fvg
/pine-patterns htf
/pine-patterns alert-json

Available Patterns:

/pine-to-python¶

Convert PineScript indicator logic to Python for backtest parity.

/pine-to-python STRAT_02.pine
/pine-to-python STRAT_02.pine --output src/strategies/STRAT_02/

Output:

• detector.py - Main detection logic
• indicators.py - Helper functions
• tests/test_parity.py - Parity verification

/pine-learn¶

Interactive PineScript learning with tutorials and examples.

/pine-learn state-machines
/pine-learn anti-repaint
/pine-learn arrays
/pine-learn "how do I detect order blocks?"

Topics:

Python & ML¶

/backtest¶

Run comprehensive backtests with anti-bias validation.

/backtest STRAT_02 --symbol MNQ --timeframe 4h
/backtest STRAT_02 --walk-forward --monte-carlo
/backtest STRAT_02 --compare-baseline

Options:

Output:

• outputs/backtests/STRAT_02_MNQ_4h_report.md
• outputs/backtests/STRAT_02_MNQ_4h_trades.csv

/train-model¶

Train and deploy ML models for edge prediction.

/train-model STRAT_02 --symbol MNQ --timeframe 4h
/train-model STRAT_02 --model xgboost --cv 5
/train-model STRAT_02 --deploy

Options:

Output:

• outputs/models/STRAT_02/model_v3.pkl
• outputs/models/STRAT_02/model_report.md

/rsi-price-engineering¶

RSI-to-Price reverse engineering and predictive algorithms.

/rsi-price-engineering reversal-math
/rsi-price-engineering price-attractor
/rsi-price-engineering rqf-integration

Algorithms:

Meta Skills¶

/skill-create¶

Create new Claude Code skills with quality scoring.

/skill-create deploy-strategy "Deploy validated strategy to paper/live"
/skill-create market-scanner "Scan for RQF setups"

Process:

1. Intent analysis
2. Architecture design
3. Write skill
4. Quality scoring (0-100)
5. Self-reflection & iteration
6. Create files

/skills-list¶

Quick reference for all available skills.

/skills-list

Skill Comparison¶