OpenClaw Prediction Market Toolkit 📈⚖️

A comprehensive toolkit for accessing prediction market data and finding cross-platform arbitrage opportunities for autonomous agents. Powered by AIsa.

This repository contains skills and Python clients that allow AI agents and developers to interact with prediction markets like Polymarket and Kalshi using a single, unified API.

📁 Included Files

Skill

skill.md: The OpenClaw skill definition for cross-platform prediction-market arbitrage lookups.

Scripts

script/prediction_market_client.py: CLI and Python wrapper for Polymarket, Kalshi, and cross-platform market-matching endpoints.
script/arbitrage_finder.py: CLI that scans matching markets, compares prices, calculates spreads, and checks liquidity.

Setup

Both the skill and the Python scripts require an AIsa API key.

export AISA_API_KEY="your-key-here"

Using the Python Scripts

Run the scripts from the repository root with the script/ path prefix.

Prediction Market Client

# Search for open election markets
python script/prediction_market_client.py polymarket markets --search "election" --status open

# Get current price for a specific token
python script/prediction_market_client.py polymarket price <token_id>

# Get wallet portfolio and PnL
python script/prediction_market_client.py polymarket positions <wallet_address>
python script/prediction_market_client.py polymarket pnl <wallet_address> --granularity day

# Search for Fed rate markets
python script/prediction_market_client.py kalshi markets --search "fed rate" --status open

# Get orderbook for a market
python script/prediction_market_client.py kalshi orderbooks --ticker <ticker>

Arbitrage Finder

# Scan all NBA games on March 30, 2025, showing only spreads > 2% and liquidity > $500
python script/arbitrage_finder.py scan nba --date 2025-03-30 --min-spread 2.0 --min-liquidity 500

# Check a specific event using its Polymarket slug
python script/arbitrage_finder.py match --polymarket-slug nba-lakers-vs-celtics

# Or using its Kalshi ticker
python script/arbitrage_finder.py match --kalshi-ticker KXNBA-25-LAL-BOS

Using the Skill

Load skill.md into the agent context.

Important details:

The repository contains skill.md, not SKILL.md.
The cURL examples in skill.md use product placeholders such as {token_id} and {market_ticker} by design.
Before executing any curl command, the runner must verify that every {...} placeholder has been replaced with a concrete value.
If a final command still contains {token_id} or any other brace placeholder, do not execute it. Fail fast and report a missing-parameter error.

ID Lookup Workflow

Most endpoints require IDs fetched from a markets endpoint first:

token_id: from side_a.id or side_b.id in /polymarket/markets
condition_id: from condition_id in /polymarket/markets
market_ticker: from market_ticker in /kalshi/markets

Pricing

| API Operation | Cost | |---------------|------| | Prediction market read query | $0.01 |

Every response includes usage.cost and usage.credits_remaining.

Documentation

For complete endpoint documentation, see the AIsa API Reference: https://docs.aisa.one/reference/