Skip to content
Get packs

Strategy Development Manager

Turns a research paper into a registered, backtested, decay-monitored trading factor or strategy

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

What it does

This skill guides an AI agent through the full path from an academic paper or research report to a tested trading factor or strategy inside the Vibe-Trading system. The agent reads the paper, pulls out formulas and rules, writes code that implements them, runs a backtest, checks results against statistical thresholds, and keeps records of each artifact so its performance can be monitored for decay over time. It delegates the heavy work to other tools in the system rather than doing it itself.

Great for

  • Quantitative researchers who want to test published factors on their own market data
  • Analysts who need a repeatable record of which research ideas were tested and how they performed
  • Teams running periodic checks to see whether an existing factor's edge is fading

Skip it if

  • Anyone expecting live trading advice or guaranteed returns: backtest results on historical data do not predict future performance and the thresholds used are simple rules of thumb
  • Users without access to market data and the Vibe-Trading backtest tools, since most steps depend on them
  • Users who want to trade real money automatically without reviewing the generated code, because the skill writes and executes strategy code that may contain errors or lookahead bias

Try asking

  • “Read this paper on momentum factors and build a backtested factor from it”
  • “Check whether my momentum_12_1 factor is still working and flag any decay”
  • “List my registered factors and show their status”
  • “Disable the value factor I registered last month”
Step-by-step guide: how to use Strategy Development ManagerInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 9 original files, the MIT licence and an attribution note.

    strategy-dev-manager.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

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

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip strategy-dev-manager.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

  • Review the generated signal_engine.py before running the backtest, since the agent writes it from a template and the TODO sections must be filled correctly to avoid lookahead bias
  • Check the OCR quality flag after ingesting a paper and manually verify formulas when text density is low
  • Set bench_type to initial or periodic in config.json, otherwise scheduled decay monitoring will ignore the run

What's inside

  • examples.md
  • references/decay_thresholds.md
  • references/scheduled_decay_scan.md
  • references/strategy_extraction_guide.md
  • references/strategy_metrics.md
  • SKILL.md
  • templates/decay_report.md
  • templates/factor_signal_engine.py
  • templates/strategy_signal_engine.py

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.