Skip to main content
skill CX Codex

backtest · starchild-ai-agent — Claude Code skill: what it does, how to install

backtest runs historical simulations for crypto trading ideas to report actual returns, drawdowns, and performance metrics.

no license 217 installs ★ 33 updated 2026-10-22

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add Starchild-ai-agent/official-skills --skill backtest -a claude-code
Open on GitHub

What backtest does

The skill converts entry and exit rules into runnable Python scripts. Before generating code, the assistant clarifies target assets, time horizons, and triggers. Code runs through existing tools in scripts/ or generates new ones, producing performance charts and summaries in output/.

It handles two core execution modes: fast vectorized passes for basic signal tests and bar-by-bar loops for orders with stop-loss or dynamic position sizing. Logic automatically shifts indicators by one step to counter look-ahead bias and factors in a baseline 0.1% trade fee.

Without a valid COINGECKO_API_KEY and the pandas, numpy, and mplfinance Python packages, data retrieval and execution will fail.

What it helps with

  • Test moving average or RSI indicators against historical crypto prices
  • Calculate portfolio drawdown, total return, and Sharpe ratios
  • Simulate order execution with trading fees and realistic bar delays
  • Generate equity curve charts and summary visuals

Example prompts for the agent

  • Backtest an EMA crossover strategy on BTC using historical hourly data
  • Run an RSI mean-reversion test on ETH and calculate max drawdown with fees
  • Evaluate a walk-forward MACD setup and save equity charts to output

How to install backtest

1Skills CLI (Vercel)

npx skills add Starchild-ai-agent/official-skills --skill backtest -a claude-code

2Manually: put the files in place

~/.claude/skills/backtest/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/backtest/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about backtest

How to install backtest in Claude Code?

Run npx skills add Starchild-ai-agent/official-skills --skill backtest -a claude-code.

What does the backtest skill do?

It builds and executes backtesting routines for crypto strategies, outputting drawdown numbers, return rates, and equity charts.

What prerequisites are required for backtesting?

You need COINGECKO_API_KEY set in your environment alongside pandas, numpy, and mplfinance installed via pip.

Similar skills

All skills for Finance and business →

backtest by other authors comes first.

backtest marketcalls backtest builds a Python script to test a trading strategy against a chosen symbol. It can report performance, compare a benchmark, and plot results. CX Codex +1 3,500installs quant-analyst rmyndharis quant-analyst — a skill for Claude Code and Codex designed for financial modeling and backtesting. Handles portfolio optimization, risk metrics, and time series analysis. CX Codex +1 123installs backtest-lab Shubham Kumar The backtest-lab Claude Code skill tests rule-based swing strategies on NSE and BSE daily data. It activates for requests to backtest a rule, inspect drawdowns, or account for trading charges. CX Codex +1 ★ 69repo stars backtest-expert tradermonty backtest-expert guides systematic trading-strategy tests with slippage, bias checks, parameter stress tests, and result review. CX Codex +1 4,700installs correlation-analysis agiprolabs correlation-analysis computes correlations between assets: rolling windows, hierarchical clustering, tail dependence. Useful for diversification checks and pairs trading. CX Codex +1 530installs vectorbt agiprolabs vectorbt runs vectorized backtests for trading strategies using NumPy and pandas array operations. It executes parameter grid searches and analyzes portfolio risk. CX Codex +1 566installs backtesting-trading-strategies Jeremy Longshore backtesting-trading-strategies — a skill that simulates trading strategies on historical data and reports risk metrics. Use it when a signal needs checking or two approaches need comparing inside the agent. CX Codex +1 4,800installs

From the same set66

The whole set →

Project sponsors

Companies that keep this analytics open