Skip to content
Get packs

Strategy Generate

Writes, backtests and evaluates quantitative trading strategies from a plain-language description

Source Download skill
Repository
HKUDS/Vibe-Trading
Stars
37,433
Licence
MIT
Files
2 files

What it does

When installed, the agent can turn a description of a trading idea into a Python signal engine, run it through a built-in backtest tool, and report metrics such as returns and risk, plus whether the results pass its review criteria. It also asks for missing details like the instrument or date range instead of guessing, and it is instructed to report poor results honestly rather than tuning the strategy until it looks good. It can also adapt a registered strategy to a different set of instruments while keeping the original signal logic.

Great for

  • Quant researchers or students who want to prototype a trading idea quickly and test it on historical data
  • Developers building a backtesting workflow who want a structured, contract-based strategy format
  • Analysts who want an honest first-pass evaluation of a strategy idea before investing more time

Skip it if

  • Live trading or investment decisions: the skill only backtests on historical data, and the files say a successful backtest is not independent approval, so past results do not show future performance
  • Users who need a production trading system: the skill produces research code that has not been validated beyond the stated checks
  • Non-technical users without a way to review generated code, since it writes and executes Python that can contain logic errors

Try asking

  • “Backtest a 50/200-day moving average crossover on AAPL and MSFT over the last 10 years”
  • “Create a momentum strategy that holds the top 10 stocks by 12-month return and rebalances monthly”
  • “Adapt a registered strategy to a different set of instruments with a specific date window”
  • “Check whether my mean-reversion idea still works on a 2018 to 2024 window with 0.1% commission”
Step-by-step guide: how to use Strategy GenerateInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 2 original files, the MIT licence and an attribution note.

    strategy-generate.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip strategy-generate.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip strategy-generate.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give explicit instrument codes, date range and entry/exit rules; the skill asks questions rather than guessing when these are missing
  • Specify the commission and initial capital if they differ from the defaults of 0.1% and 1,000,000
  • Treat backtest results as research output: check the evidence and cost warnings, and do not read a good Sharpe ratio as proof the strategy will work

What's inside

  • examples.md
  • SKILL.md

Redistributed unmodified from HKUDS/Vibe-Trading under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.