Prediction Market Monitor

prediction-market-monitor

You pass a Kalshi series or shortcut. The skill pulls open markets, groups by event, and reports implied probabilities. When the event is a laddered strike-type (Fed funds level, CPI reading), it derives the cross-strike probability distribution from adjacent-threshold differences and reports the modal outcome and expected value.

Motivated by 2025-26 growth of prediction markets as leading indicators post-2024 election validation. Kalshi Fed-decision contracts now trade meaningful volume; contract prices often reflect policy expectations before consensus surveys catch up.

When to invoke

• "What's the Kalshi-implied Fed decision at the next meeting?"
• Comparing market-implied CPI to Bloomberg / Reuters consensus
• Cross-referencing macro-print bets against your positioning
• The user says "Kalshi", "prediction market", "implied Fed", "fed funds futures alternative"

What you need

• Internet access (Kalshi's public API; no Massive key needed)

Optional:

• --series (shortcut like fed, cpi, nfp, gdp, or a raw Kalshi series ticker like KXFED, KXCPI)
• --keyword (client-side filter on title / ticker)
• --event-ticker (pin a specific event, e.g. KXFED-27APR)
• --max-events (default 5)

What you get back

Layer 1: JSON. Per event: event_ticker, title, close_time, markets (each with implied_probability, bid, ask, last, volume, open interest, floor_strike). When laddered: implied_distribution.buckets with p_in_bucket and cumulative_p_above_lower, plus modal_bucket and expected_value.

Layer 2: rendered note. Per event: title + close time, modal outcome + expected value line, bucket distribution with ASCII bars.

How it works

1. Query Kalshi /trade-api/v2/markets with the series filter. Paginate up to 3 pages (600 markets max).
2. Group markets by event_ticker. A single event ("KXFED-27APR") typically contains 15-20 laddered strikes.
3. For each event, sort by floor_strike and derive the implied distribution: P(rate in [lower, upper)) = P(above lower) - P(above upper).
4. Report modal outcome (highest-probability bucket) and expected value (probability-weighted midpoint).

Series shortcuts

• fed: KXFED (fed funds level after meeting)
• fed_decision: KXFEDDECISION (rate change at meeting)
• cpi: KXCPI (m/m)
• core_cpi: KXCORECPI
• cpi_yoy: KXCPIYOY
• nfp: KXNFP
• unemployment: KXUNEMP
• gdp: KXGDP
• jobless_claims: KXICSA
• recession: KXRECESSIONYEAR
• spx_close: KXSPX
• btc_close: KXBTCD

Foundations used

• None. Kalshi public API only.

Endpoints used

• GET https://api.elections.kalshi.com/trade-api/v2/markets Public read-only, no auth.

Doesn't handle (yet)

• Polymarket integration. Kalshi only. Polymarket has more political / cultural markets; Kalshi has more macro. Adding a --source polymarket toggle would be a clean extension.
• Time series of implied probability. Snapshot only. A rolling history would show when the market moved.
• Cross-reference to survey consensus. Would need a data partnership with Bloomberg / Reuters or an FOMC dot-plot lookup.
• CFTC-regulated futures cross-check. SOFR/Fed funds futures implied path from CME data would be a nice comparator.

These are clean PR extensions.