molt-pmxt_skill

Overview

This skill grants real-time access to prediction markets (Polymarket, Kalshi, Limitless) for rapid fact-checking, probability analysis, and order execution. It combines discovery, precise quote retrieval, and execution capabilities so an agent can find relevant markets, report implied probabilities, and place trades when explicitly instructed. The skill is built for live market workflows and includes safeguards to prevent unintended orders.

How this skill works

The skill exposes three core actions: pmxt_search to discover active markets by keyword, pmxt_quote to fetch exact yes/no prices and convert them to implied probabilities, and pmxt_order to execute live orders when the user confirms amount and outcome. It searches exchanges concurrently, synthesizes quote data into readable probabilities, and checks for cross-exchange arbitrage opportunities to notify users. Trading requires configured API keys and private keys in the runtime environment.

When to use it

• You need an empirical market-implied probability for a real-world event.
• You want to find active, high-volume markets related to a topic for fact-checking.
• You plan to place a real-money bet and have explicitly confirmed amount and outcome.
• You need to detect or be alerted to cross-exchange arbitrage discrepancies.
• You want to monitor market sentiment or trending markets for research or reporting.

Best practices

• Use broad keyword queries (not natural-language questions) for pmxt_search to maximize relevant results.
• Always call pmxt_quote after pmxt_search before interpreting probabilities; present prices as percentages.
• Never execute pmxt_order without explicit user confirmation specifying amount, outcome, and exchange.
• Keep API and private keys secure in environment variables; do not expose keys in conversation.
• Silently check for arbitrage after quotes and alert users immediately if opportunities exist.

Example use cases

• User asks: 'What chance does Market X assign to Event Y?' — run pmxt_search, pmxt_quote, then return the implied percentage.
• User requests trending markets on 'crypto' — use pmxt_search focused on high-volume markets and summarize.
• User says 'Buy $50 of Yes on market ABC' — verify explicit intent and details, then call pmxt_order to execute.
• Analyst monitoring for arbitrage — run pmxt_quote across exchanges and notify when price sums imply risk-free profit.
• Reporter fact-checking a claim — use market prices to report the market-implied likelihood rather than model-based estimates.