vectorbt-expert — Claude Code skill
A skill for testing trading strategies with VectorBT, from signal generation and historical data to performance charts and benchmark comparison.
npx skills add marketcalls/vectorbt-backtesting-skills --skill vectorbt-expert -a claude-codeWhat vectorbt-expert does
vectorbt-expert guides an agent through Python backtesting with vectorbt, pandas, numpy, and plotly. It covers historical data retrieval, entry and exit signals, position sizing, market fees, portfolio metrics, drawdowns, equity curves, and trade analysis.
The default indicator layer is openalgo.ta, including exrem(), crossover(), crossunder(), and flip(). The skill requires a strategy-versus-benchmark table after each backtest and includes guidance for OpenAlgo, DuckDB, yfinance, CCXT, parameter searches, and walk-forward analysis.
Scripts belong in backtesting/{strategy_name}/ directories. API settings are read from the project-root .env through python-dotenv and find_dotenv(). Python and the vectorbt, pandas, numpy, and plotly packages are required for these scripts to run.
What it helps with
- backtest a trading strategy
- generate entry and exit signals
- compare a strategy with its benchmark
- inspect drawdown and position sizing
- search for better strategy parameters
Example prompts for the agent
- Backtest an EMA crossover strategy with OpenAlgo data and compare it with the benchmark.
- Load historical prices from DuckDB, create entry and exit signals, then review the trades.
- Test the strategy parameters in
vectorbtand plot the equity curve and drawdown.
How to install vectorbt-expert
1Skills CLI (Vercel)
npx skills add marketcalls/vectorbt-backtesting-skills --skill vectorbt-expert -a claude-code2Manually: put the files in place
~/.claude/skills/vectorbt-expert/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/vectorbt-expert/
1Skills CLI (Vercel)
npx skills add marketcalls/vectorbt-backtesting-skills --skill vectorbt-expert -a codex2Manually: put the files in place
.agents/skills/vectorbt-expert/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add marketcalls/vectorbt-backtesting-skills --skill vectorbt-expert -a cursor2Manually: put the files in place
.cursor/skills/vectorbt-expert/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install marketcalls/vectorbt-backtesting-skills vectorbt-expert2Manually: put the files in place
.github/skills/vectorbt-expert/SKILL.md1Skills CLI (Vercel)
npx skills add marketcalls/vectorbt-backtesting-skills --skill vectorbt-expert -a gemini-cli2Platform CLI
gemini skills install https://github.com/marketcalls/vectorbt-backtesting-skills --path .claude/skills/vectorbt-expertManually: put the files in place
.devin/skills/vectorbt-expert/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add marketcalls/vectorbt-backtesting-skills --skill vectorbt-expert -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about vectorbt-expert
How do I install vectorbt-expert in Claude Code?
Run npx skills add marketcalls/vectorbt-backtesting-skills --skill vectorbt-expert -a claude-code. The skill file will be at ~/.claude/skills/vectorbt-expert/SKILL.md.
What does vectorbt-expert do?
It helps the agent build and inspect VectorBT backtests, fetch data, create signals, configure positions, calculate metrics, and draw charts. Each backtest also gets a benchmark comparison and a plain-language report.
Which indicator tools does the skill use?
It uses openalgo.ta as the default indicator layer, including signal-cleaning helpers. TA-Lib is used only when explicitly requested, while VectorBT built-in indicators are excluded.
