arb_polyweather

This bot uses an AI agent to make trading decisions on Polymarket weather markets, making it distinct from a simple rules-based or ML-model bot. Its approach likely involves gathering live weather data (e.g., from Open-Meteo), feeding it into an LLM (e.g., Claude) which is prompted to reason about the likely outcome (e.g., 'Will it be above 100°F in Phoenix?'), and then comparing that reasoning with the market-implied probability to execute trades. The 26 MB size suggests it might contain a JSON of prompts, response caches, or included dependencies. Code quality is not obvious from size alone, but the novel approach makes it a useful case study. The risk here is the 'AI agent' part; LLM hallucinations could produce flawed reasoning, so robust prompt engineering and output validation are critical. Malware risk is low. Who should use it: ML engineers wanting to explore LLM-driven trading or those looking to blend historical weather data with language model reasoning. The bot represents a step beyond just 'data in, signal out' to a more complex, agent-based trading paradigm.