Add comprehensive Kilo Code agent skills suite for autonomous strategy development

[itsXactlY]

Kilo Code Agent Skills Suite for BTQuant

Overview

This PR adds an extensive suite of agent skills for Kilo Code CLI, enabling fully autonomous strategy development, testing, optimization, and evaluation workflows in BTQuant.

What's Included

Generic Skills (All Modes) - 11 Skills

1. strategy-development - Strategy architecture, parameters, signal generation, risk management
2. strategy-backtesting - Backtest setup, data sourcing, metrics, validation
3. strategy-optimization - Grid search, Bayesian optimization, robustness testing
4. strategy-evaluation - Performance metrics, evaluation framework, production criteria
5. data-management - Data sourcing, validation, preparation, feature engineering
6. exchange-integration - CCXT setup, order management, websockets, rate limiting
7. ml-models - Classification/regression models, feature engineering, training best practices
8. autonomous-workflows - 6-stage development pipeline, continuous monitoring, CI/CD
9. code-quality - PEP 8, testing, documentation, security, error handling
10. performance-analysis - Return analysis, risk metrics, visualization, comparisons
11. documentation - README templates, changelogs, API docs, performance reports

Code Mode Skills - 2 Specialized Skills

1. testing-strategies - Unit tests, integration tests, edge cases, parametrized tests
2. advanced-debugging - Signal tracing, profiling, memory tracking, logging

Key Features

Autonomous Development Pipeline

Skills enable agents to:

Generate Strategies → Validate Code → Backtest → Optimize → Evaluate → Document

Multi-Stage Workflow Support

• Stage 1: Strategy generation with parameter configs
• Stage 2: Syntax validation and linting
• Stage 3: Comprehensive backtesting on multiple assets
• Stage 4: Bayesian parameter optimization
• Stage 5: Multi-metric evaluation against production criteria
• Stage 6: Auto-generated documentation

Continuous Monitoring

• Daily performance tracking
• Automatic regime detection
• Triggered retraining when performance degrades > 25%
• Comprehensive alert system

Advanced Capabilities

• ML Integration: Classification, regression, time-series validation
• Exchange Integration: CCXT support, 400+ exchanges, testnet trading
• Data Validation: OHLCV integrity checks, outlier detection
• Performance Analysis: Sharpe, Sortino, Calmar, profit factor calculations
• Production Ready: Comprehensive testing and quality standards

Usage

Quick Start

cd /path/to/BTQuant
kilocode

# Tasks automatically leverage skills:
> "Generate 3 new MA crossover strategy variants"
> "Run comprehensive backtest suite"
> "Optimize with Bayesian search"
> "Evaluate against production criteria"

Code Mode

/mode code

# For testing and debugging:
> "Add unit tests for signal generation"
> "Debug why backtest performance degrades"

Autonomous CI/CD

echo "Run full development pipeline" | kilocode --auto --timeout 600

Directory Structure

.kilocode/
├── skills/                          # Generic skills (all modes)
│   ├── strategy-development/
│   ├── strategy-backtesting/
│   ├── strategy-optimization/
│   ├── strategy-evaluation/
│   ├── data-management/
│   ├── exchange-integration/
│   ├── ml-models/
│   ├── autonomous-workflows/
│   ├── code-quality/
│   ├── performance-analysis/
│   └── documentation/
├── skills-code/                     # Code mode specific
│   ├── testing-strategies/
│   └── advanced-debugging/
└── SKILLS_README.md                 # Comprehensive documentation

Example Workflows

Workflow 1: Strategy Development & Optimization

1. Agent creates new strategy (strategy-development)
2. Runs backtest (strategy-backtesting)
3. Optimizes parameters (strategy-optimization)
4. Evaluates viability (strategy-evaluation)
5. Generates documentation (documentation)

Workflow 2: ML Strategy Development

1. Design ML model (ml-models)
2. Engineer features (data-management)
3. Train with time-series validation (ml-models)
4. Backtest (strategy-backtesting)
5. Optimize hyperparameters (strategy-optimization)
6. Analyze performance (performance-analysis)

Workflow 3: Production Deployment

1. Meet evaluation thresholds (strategy-evaluation)
2. Pass code quality review (code-quality)
3. Complete performance analysis (performance-analysis)
4. Create documentation (documentation)
5. Set up continuous monitoring (autonomous-workflows)

Benefits

✅ Fully Autonomous: Zero human intervention required for standard workflows
✅ Comprehensive: Covers entire development lifecycle
✅ Production-Ready: Includes validation, testing, and monitoring
✅ Extensible: Easy to add new skills or customize existing ones
✅ Well-Documented: Each skill includes practical code examples
✅ Mode-Specific: Generic skills available everywhere, code skills in code mode
✅ Best Practices: Incorporates industry standards and BTQuant conventions
✅ Risk Management: Built-in production criteria and validation

Integration

These skills work seamlessly with:

• Kilo Code CLI interactive and autonomous modes
• BTQuant's CCXT integration (400+ exchanges)
• MSSQL data backend
• ML frameworks (sklearn, PyTorch)
• Python testing frameworks (pytest, unittest)

Next Steps

1. Start interactive session: kilocode
2. Try a simple task: "Create a simple MA crossover strategy"
3. Scale up to complex workflows
4. Customize skills for your specific needs

References

• Kilo Code Documentation: https://kilo.ai/docs/cli
• BTQuant Repository: https://github.com/itsXactlY/BTQuant
• Agent Skills Specification: https://github.com/kilo-agents/skills

---

Status: Ready for autonomous agent deployment
Total Skills: 13 (11 generic + 2 code-mode)
Lines of Documentation: 3,500+
Code Examples: 150+