Agent Skills
› NeverSight/learn-skills.dev
› quant-analyst
quant-analyst
GitHub专注于量化金融、算法交易及金融数据分析。适用于构建交易策略、回测框架、风险模型（VaR/CVaR）、组合优化及衍生品定价，利用Python进行统计建模与机器学习应用。
Trigger Scenarios
构建算法交易策略或回测框架
执行金融时间序列的统计分析
实现风险模型如VaR和Greeks计算
创建投资组合优化算法
开发衍生品定量定价模型
Install
npx skills add NeverSight/learn-skills.dev --skill quant-analyst -g -y
SKILL.md
Frontmatter
{
"name": "quant-analyst",
"description": "Expert in quantitative finance, algorithmic trading, and financial data analysis using Python (Pandas\/NumPy), statistical modeling, and machine learning."
}
Quantitative Analyst
Purpose
Provides expertise in quantitative finance, algorithmic trading strategies, and financial data analysis. Specializes in statistical modeling, risk analytics, and building data-driven trading systems using Python scientific computing stack.
When to Use
- Building algorithmic trading strategies or backtesting frameworks
- Performing statistical analysis on financial time series data
- Implementing risk models (VaR, CVaR, Greeks calculations)
- Creating portfolio optimization algorithms
- Developing quantitative pricing models for derivatives
- Analyzing market microstructure and order book dynamics
- Building factor models for asset returns
- Implementing Monte Carlo simulations for financial instruments
Quick Start
Invoke this skill when:
- Building algorithmic trading strategies or backtesting frameworks
- Performing statistical analysis on financial time series data
- Implementing risk models (VaR, CVaR, Greeks calculations)
- Creating portfolio optimization algorithms
- Developing quantitative pricing models for derivatives
Do NOT invoke when:
- Building general web applications → use fullstack-developer
- Creating data visualizations without financial context → use data-analyst
- Implementing payment processing → use payment-integration
- Building generic ML models → use ml-engineer
Decision Framework
Financial Analysis Task?
├── Trading Strategy → Backtesting framework + signal generation
├── Risk Management → VaR/CVaR models + stress testing
├── Portfolio Optimization → Mean-variance, Black-Litterman, risk parity
├── Derivatives Pricing → Monte Carlo, finite difference, analytical
└── Time Series Analysis → ARIMA, GARCH, cointegration tests
Core Workflows
1. Algorithmic Trading Strategy Development
- Define trading hypothesis and signal generation logic
- Implement strategy using vectorized Pandas operations
- Build backtesting engine with realistic execution simulation
- Calculate performance metrics (Sharpe, Sortino, max drawdown)
- Perform walk-forward optimization to avoid overfitting
- Implement live trading hooks with proper risk controls
2. Risk Model Implementation
- Gather historical price/returns data
- Select appropriate risk metric (VaR, CVaR, Greeks)
- Implement calculation using parametric, historical, or Monte Carlo methods
- Validate model with backtesting and stress scenarios
- Build monitoring dashboard for real-time risk exposure
3. Portfolio Optimization
- Define investment universe and constraints
- Calculate expected returns and covariance matrix
- Implement optimization (scipy.optimize or cvxpy)
- Apply regularization to prevent concentration
- Rebalance periodically with transaction cost consideration
Best Practices
- Use vectorized NumPy/Pandas operations for performance on large datasets
- Always account for transaction costs, slippage, and market impact in backtests
- Implement proper cross-validation (walk-forward) to prevent lookahead bias
- Use log returns for statistical properties, simple returns for aggregation
- Store financial data with timezone-aware timestamps (UTC preferred)
- Validate models with out-of-sample testing before deployment
Anti-Patterns
- Overfitting to historical data → Use walk-forward validation and regularization
- Ignoring transaction costs → Include realistic costs in all backtests
- Using future data in signals → Ensure strict point-in-time correctness
- Assuming normal distributions → Use fat-tailed distributions for risk models
- Hardcoding market assumptions → Parameterize and stress test assumptions
Version History
- e0220ca Current 2026-07-05 21:17
