← ClaudeAtlas

polymarket-eventslisted

Manually refresh prediction-market implied probabilities for macro + crypto + geopolitical events from Polymarket's Gamma public-search API. No auth, no rate limit at retail-research scale. Returns the current probability + resolution date for tracked event categories (Fed rate cuts, recession, inflation, BTC/ETH price targets, geopolitical risk). Macro-confluence leg of the §4 sentiment stack — feeds the dashboard's Event Probabilities panel and the §5 halt-window doctrine. Manual by design — no automatic refresh, no cron.
kensongan-prog/trading-advisor · ★ 1 · API & Backend · score 62
Install: claude install-skill kensongan-prog/trading-advisor
# Polymarket Events Skill ## When to use Trigger this skill when the user says: - "Refresh prediction markets" / "Refresh Polymarket" - "What does Polymarket think about Fed cuts?" - "What's the recession probability?" - "Show me the event probabilities" Do NOT trigger for: - Individual ticker sentiment (use `reddit-sentiment` + `stocktwits-sentiment` + `sentiment-cache`) - Macro fundamentals (use `macro-rates` for FRED data — actual rates, not implied) - Scheduled-event halt windows (use `macro-calendar` — that's the event *dates*, this skill is the implied probability of *outcomes*) ## Why this exists Polymarket is the only large-scale liquid prediction market open to the public. It aggregates the *money-weighted consensus* of speculators on: - **Fed policy** (rate cuts/hikes, FOMC outcomes) - **Macro outcomes** (recession, inflation prints, unemployment) - **Crypto price targets** (BTC/ETH at specific dates) - **Geopolitics** (tail-risk events) These signals are *different* from §4 retail forum sentiment (cheap talk) — Polymarket participants are putting cash on outcomes, so the implied probabilities are less gameable. The doctrine treats Polymarket as a **macro confluence signal**, not a contrarian filter: - **Aligned** (Polymarket agrees with our macro thesis) → confluence boost - **Diverged** (Polymarket disagrees) → reconsider thesis, don't auto-fade - **Tight uncertainty** (probabilities clustered near 50%) → respect the noise, don't overcommit §5 halt-window