LumiBot Trading Skills

Free trading strategy skills for Claude Code, Codex and Cursor, built on LumiBot.

The guides and catalog.json are generated from the same maintained skill source used by BotSpot's authoring agent. They are published together after privacy and parity checks; there is no separately edited public version. Installed native skills keep the version you installed until you update the package. Compatible BotSpot MCP server versions read the published catalog and refresh its content automatically; older servers retain their bundled version until their normal update.

Ask your AI to build a trading bot, and these skills teach it how to do it right. It learns how to write a LumiBot strategy, backtest it, read the results, and avoid the mistakes that make backtests lie. They cover stocks, ETFs, options, crypto, futures and AI agents that trade.

LumiBot is the open-source Python trading library from Lumiwealth. Docs: https://lumibot.lumiwealth.com

Install in one command

Claude Code

/plugin marketplace add Lumiwealth/lumibot-trading-skills
/plugin install lumibot-trading-skills@lumiwealth

Or from your shell:

claude plugin marketplace add Lumiwealth/lumibot-trading-skills
claude plugin install lumibot-trading-skills@lumiwealth

Codex, Cursor, Claude Code and 70+ other agents (with the skills CLI, Node 22 or newer):

npx skills add Lumiwealth/lumibot-trading-skills

Add -g to install for every project, or -a codex / -a cursor / -a claude-code to pick one agent.

Codex by hand. Codex reads personal skills from ~/.agents/skills:

git clone https://github.com/Lumiwealth/lumibot-trading-skills.git
mkdir -p ~/.agents/skills
cp -R lumibot-trading-skills/skills/* ~/.agents/skills/

Restart Codex if the skills do not show up right away.

The skills

Skill	What it helps with
lumibot-strategy-building	Start here. Writing, fixing and backtesting any LumiBot strategy, plus the checks that catch bad backtests.
lumibot-backtesting	Diagnose failed or no-trade backtests from logs, fills and artifacts; use local LumiBot or BotSpot MCP.
lumibot-api-reference	Exact LumiBot method names, lifecycle methods and working code examples.
stock-etf-strategies	Stocks and ETFs: moving averages, breakouts, mean reversion, rotation and rebalancing.
options-strategies	Options: covered calls, the wheel, spreads, iron condors, straddles and more.
lumibot-indicators	Indicators like SMA, RSI, MACD and Bollinger Bands, and putting them on the backtest chart.
lumibot-ai-agents	Strategies where AI agents do the research and one AI agent places the trades.
fast-lumibot-strategies	Making slow strategies and big scanners fast, and handling orders quickly and safely.
using-botspot-mcp	Using the free BotSpot MCP to build, backtest and run strategies in the cloud.

Each skill folder has a SKILL.md and a references/ folder with deeper notes. Your AI opens the reference files only when it needs them. Some skills mention a skills tool for loading references. That is how the BotSpot MCP serves them. Here, the files are already on your computer in references/.

Quick start

1. Install the skills with one of the commands above.
2. Ask your AI: "Build me a moving average strategy for SPY and backtest it."
3. Run it with LumiBot: pip install lumibot, then run the strategy file your AI wrote. The LumiBot docs show how to backtest and how to connect a broker.

Connect to hosted backtests

Do not want to set up Python and data? Connect the BotSpot MCP. It is free to connect, and it creates a free BotSpot account for you. Your AI can then use strategy-building skills, research data and hosted backtest tools. Hosted compute and heavier usage depend on your plan.

Claude Code:

claude mcp add --transport http botspot https://mcp.botspot.trade/mcp

Other apps that support MCP can use the same address: https://mcp.botspot.trade/mcp. Learn more at botspot.trade.

Running a bot with your own broker always needs your approval on botspot.trade first. Nothing trades without you saying yes.

Good to know

• These skills teach an AI how to write code. They are not investment advice.
• A backtest shows what would have happened in the past. It does not promise anything about the future.
• Trading can lose money. Test with a paper account before you use real money.

Check the layout

python3 tests/test_skills_layout.py
python3 -m unittest discover -s tests

License

MIT. See LICENSE.