aabrole

Claude Trading Desk — Development skill for Claude Code

Development community

Build, honestly backtest, paper trade and self-host algorithmic trading bots with Claude Code.

How to install Claude Trading Desk

This entry records only its repository, not the path inside it, so there is no exact command to give. Open aabrole/claude-trading-desk and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Claude Trading Desk does

Build, honestly backtest, paper trade and self-host algorithmic trading bots with Claude Code. Free stack end to end.

Alternatives in Development

Skill Decision Support — Present options with trade-offs for informed decision-making — use when choosing between approaches 2.8k ★
02 One Shot Trade — One-Shot Onboarding Prompt 584 ★
Ship Studio — The professional studio for modern development — build it, ship it, host it without leaving the app 249 ★

README

Trading Desk

A Claude Code plugin for building algorithmic trading bots, testing them honestly, paper trading them, and hosting them 24/7 for nothing.

Everything in this stack has a free tier. There is no paid step anywhere.

Install

/plugin marketplace add aabrole/claude-trading-desk
/plugin install trading-desk@trading-desk

Then, in the directory you want the workspace:

/trading-desk-init

What you get

**Six skills** Claude loads when the work calls for them:

Skill	What it carries
`trading-desk-method`	The testing discipline, and ten named bugs that each made results look better than they were
`trading-desk-strategy`	The strategy contract, and the causality rules that keep a backtest honest
`trading-desk-jev`	Wiring a decision model as a filter, and how to tell whether it added anything
`trading-desk-paper`	Going live on Alpaca paper: replay validation, kill switches, free-data sharp edges
`trading-desk-dashboard`	The live desk, server-sent events, animated backtest curves
`trading-desk-oracle`	Free 24/7 hosting, capacity retries, and the two firewalls that catch everyone

**Three commands:** `/trading-desk-init`, `/trading-desk-new-strategy`, `/trading-desk-deploy`.

**A working harness** in `template/`: a portfolio backtest engine with real slippage and fees, a three-arm decision layer, a live dashboard, and a scripted deploy to a free VM.

The part that makes it worth using

Most trading repos help you find a strategy that works. This one is built to tell you when one **does not**, because that is the far more common answer and the expensive thing to get wrong.

Every strategy runs three ways over an identical candidate set:

Arm	What it is
`rules`	Take every signal. The strategy as advertised.
`gated`	Hand-written if-statements filter the signals. The control.
`jev`	A decision model filters the same signals.