kalshi-trader

Kalshi Trader Plugin

A Claude Code plugin for AI-assisted research and trading on Kalshi prediction markets.

Philosophy

We have an edge in research depth, not speed. This plugin leverages AI's ability to:

• Deeply analyze primary sources (Fed minutes, legislation, economic data)
• Synthesize information across domains
• Identify mispricings in low-volume markets where fewer participants mean more opportunity

Prerequisites

1. Claude Code installed and working
2. kalshi-cli installed and configured:

   pip install kalshi-cli
   

3. Kalshi account with API credentials configured (see kalshi-cli docs)

Installation

From GitHub

/plugin install github:austron24/kalshi-trader-plugin

Local Development

claude --plugin-dir /path/to/kalshi-trader-plugin

Commands

After installation, the following commands are available:

/kalshi-trader:alpha [criteria]

Scan for alpha opportunities and spawn parallel research agents.

/kalshi-trader:alpha                    # Default: markets closing 1-4 weeks out
/kalshi-trader:alpha fed               # Markets related to "fed"
/kalshi-trader:alpha --closes-before 2025-02-01

What it does:

1. Scans markets using kalshi CLI
2. Filters for research-friendly opportunities
3. Spawns parallel market-researcher agents
4. Reports paths to all generated research files

/kalshi-trader:creative

Run unconventional research on markets from previous /alpha scan.

/kalshi-trader:creative

What it does:

1. Identifies markets from previous /alpha in conversation
2. Spawns parallel creative-researcher agents
3. Explores non-obvious angles, contrarian views, obscure sources

/kalshi-trader:critique

Senior analyst critical review of all research.

/kalshi-trader:critique

What it does:

1. Identifies all markets with research
2. Spawns parallel senior-analyst agents
3. Critically assesses research quality, identifies flawed reasoning
4. Returns TRADE / PASS / NEEDS MORE WORK recommendations

/kalshi-trader:score

Score and rank all researched markets.

/kalshi-trader:score

What it does:

1. Spawns parallel judge agents to score each market 0-100
2. Ranks markets by opportunity quality
3. Returns top 10 opportunities

/kalshi-trader:finalize [tickers]

Final investment recommendations with trade execution.

/kalshi-trader:finalize                 # Analyze top markets from /score
/kalshi-trader:finalize TICKER1 TICKER2 # Analyze specific markets

What it does:

1. Reads ALL research for specified markets (in main context)
2. Checks portfolio status via kalshi CLI
3. Recommends positions with sizing and plain language explanations
4. Waits for user approval
5. Executes approved trades
6. Creates thesis documents for position management

Note: This command does NOT spawn subagents. It reads everything directly for full context.

/kalshi-trader:positions

Daily position review. Run this daily.

/kalshi-trader:positions

What it does:

1. Gets all open positions via kalshi CLI
2. Spawns parallel position-reviewer agents
3. Each reviews position against its thesis
4. Recommends HOLD or SELL for each
5. Presents sell recommendations for approval
6. Executes approved sells
7. Updates learnings with wins/losses

Agents

The plugin includes these specialized agents:

Agent	Purpose
market-researcher	Initial deep-dive research on market fundamentals
creative-researcher	Unconventional angles, contrarian perspectives
senior-analyst	Critical review, identifies flawed reasoning
judge	Scores markets 0-100 for ranking
position-reviewer	Reviews positions, recommends HOLD/SELL

Typical Workflow

1. /kalshi-trader:alpha              # Find opportunities
2. /kalshi-trader:creative           # Explore unconventional angles
3. /kalshi-trader:critique           # Critical review
4. /kalshi-trader:score              # Rank opportunities
5. /kalshi-trader:finalize           # Get recommendations, execute trades
6. /kalshi-trader:positions          # Daily: review and manage positions

Directory Structure

The plugin creates and uses these directories in your project:

your-project/
├── research/
│   └── markets/
│       └── <TICKER>/
│           ├── YYYY-MM-DD-initial-research.md
│           ├── YYYY-MM-DD-creative-research.md
│           ├── YYYY-MM-DD-senior-review.md
│           ├── score.txt
│           └── thesis.md              # Created by /finalize
├── state/                             # Optional
│   ├── current-state.md
│   └── session-log.md
└── learnings/                         # Optional
    ├── strategies-that-worked.md
    └── strategies-that-failed.md

Setup Templates