How to use the Portfolio Risk Metrics agent skill

What the skill makes the agent do

The skill gives the agent a working knowledge of common portfolio risk measures. Its description says it calculates VaR, CVaR, Sharpe, Sortino, and drawdown analysis, and it is meant for measuring portfolio risk, implementing risk limits, or building risk monitoring systems.

The skill also covers how these measures are used. It groups metrics into volatility (standard deviation, beta), tail risk (VaR, CVaR), drawdown (max drawdown, Calmar), and risk-adjusted performance (Sharpe, Sortino), and it lists time horizons from intraday to annual.

• Tail risk: VaR and CVaR for extreme losses
• Drawdown: max drawdown and Calmar for capital preservation
• Risk-adjusted returns: Sharpe and Sortino
• Cautions: do not rely on VaR alone, do not assume normal returns, use enough historical data

Install the skill from the ZIP download

Download the skill as a ZIP file from the page where you found this guide. Then unzip it into one of these locations, depending on how you want to use it:

• Personal skills folder: ~/.claude/skills/ (available across your projects)
• Project skills folder: .claude/skills/ inside a specific project
• Claude apps: upload the ZIP under Settings → Capabilities → Skills

Trigger the skill with natural requests

You do not need a special command. Describe the task in plain language that matches the skill's purpose: measuring portfolio risk, setting risk limits, or building a risk monitoring system. The skill is also tied to requests about risk-adjusted returns and position sizing.

The skill's files include a reference document with worked examples. The skill points the agent to that file when the main guidance is not enough, so asking for a detailed calculation can bring in those examples.

• "Explain the difference between VaR and CVaR for my portfolio returns."
• "Help me structure a drawdown analysis for this equity curve."
• "Show how to calculate the Sortino ratio with a 0% downside threshold."
• "What lookback period should I use for a 99% daily VaR?"

What you need before you start

The source lists no required software, packages, or accounts. The skill does not run scripts and does not use the network. The example code in the reference file uses Python libraries such as numpy, pandas and scipy, so you will need those if you want to run that code yourself.

• Historical return data, if you want calculations on your own portfolio
• Python with numpy, pandas and scipy, if you plan to run the reference code

Safety notes: scripts and network

The skill runs no scripts and uses no network access, according to its listing. It is written guidance and example code for the agent to use. It is not a validated risk system, and its outputs should be checked independently before anyone acts on them.

• Do not use outputs alone for live trading or position-sizing decisions.
• Do not rely on the skill for regulatory or compliance reporting. It mentions this use but does not specify rules, so an expert must verify any output.
• Check every calculation against your own data and method.

Tips for better results

Give the agent the details that change the answer: the return frequency, the lookback window, the confidence level, and any risk-free rate. The skill's own guidance says to document assumptions such as distribution and lookback.

• Use several metrics together, since no single metric captures all risk.
• Use CVaR alongside VaR, because VaR alone underestimates tail risk.
• Use rolling analysis, because risk changes over time.
• Avoid short lookbacks, which can miss regime changes.
• Include transaction costs, which affect realized risk.