Arb Bots Flood GitHub as Poly-Kalshi Spread Hunting Peaks

GitHub Intelligence — 2026-06-22

Lead Finding: Polymarket-Kalshi Arb Is the Trade of the Moment

Of the twenty repos surfaced today, seven are explicitly targeting arbitrage between Polymarket and Kalshi. That concentration is not noise. When independent developers across different geographies converge on the same strategy simultaneously, it signals one of two things: a genuine, persistent pricing inefficiency worth chasing, or a crowded trade about to compress its own edge. Based on the repo creation dates — spanning January through June 2026 — this wave has been building for six months and has not yet saturated.

The most technically credible entries in this cohort are **matthewnyc2/arbitrage** (NegRisk multi-outcome scanner, 60 passing tests, paper-trading mode, asyncio architecture) and **tduic/prediction-market** (Kelly sizing, risk-managed execution, embedded analytics dashboard). Both show hallmarks of serious engineering: test coverage, explicit risk controls, and staged deployment. Operators running their own arb infrastructure should review both repos for methodological benchmarking, even if you don't deploy the code directly.

**kmjjjj/polymarket-arbitrage-bot-btc-sol-15m** stands out for its Rust implementation targeting BTC and SOL 15-minute prediction markets. Rust-based execution implies latency sensitivity — this developer believes speed matters in these markets. If true, Python-based competitors are already behind on the fastest opportunities.

Whale Copy-Trading Tooling Emerges

**adibfarah569/PolyTrace** (created yesterday, June 21) bills itself as a "whale copy-trading terminal" currently in demo mode. This is the first repo in recent sweeps explicitly packaging on-chain wallet surveillance into a copy-trading interface. Even in demo state, its existence confirms that wallet-tracking alpha is being productized and democratized. Operators who currently hold informational advantages from manual whale monitoring should expect that edge to compress as tools like this mature.

**Dadel-1/Polymarket-CopyTrading** takes a proportional position-sizing approach to the same problem, with 3 stars and 1 fork suggesting at least some real-world uptake.

Red Flag: quipmnxailcrrgky/tradingbot

This repo demands a security callout. 98 stars, 60 forks, generic description ("Easy setup and creation of a bot"), and a randomized-string account handle that pattern-matches to throwaway or astroturfed GitHub accounts. The 34MB repo size is large for a trading bot without documented dependencies. Do not clone or run this code without full static analysis. The star and fork counts are inconsistent with zero community presence elsewhere. Treat as potentially malicious until proven otherwise — credential harvesting and wallet key exfiltration are the likely attack vectors in tooling of this profile.

**bit-nexusxtitmtdsuy/Polymarket_Bot** carries a similarly suspicious account handle pattern with 32 stars and 17 forks. Apply the same caution.

Strategy Signal: Political Market Arb Gets Disciplined

**nayrbryanGaming/arbiter** is worth a close read for operators in political markets. The repo explicitly targets no-arbitrage violations between related election markets, enforces fractional Kelly sizing, and requires backtested out-of-sample edge before any position is taken. The README philosophy — "survival over profit" — reflects a maturity level rarely seen in open-source prediction market tooling. This is the kind of framework serious operators build internally. Its public release either means the edge is already gone, or the author is building reputation. Worth watching for updates.

Quant Infrastructure: Random Forest Calibration

**BallesJr/polymarket-edge-model** packages a full systematic pipeline: Random Forest probability calibration, backtesting, and automated paper trading. Calibration models applied to Polymarket prices are a meaningful technique — raw market prices are not always well-calibrated, particularly in lower-liquidity markets. If the backtesting methodology is sound, this represents legitimate alpha-generation infrastructure.

Bottom Line

The arb-bot arms race between Polymarket and Kalshi is real and accelerating. Execution speed, calibration quality, and risk discipline are the differentiators. Two repos today show genuine engineering rigor. Two others show indicators of malicious intent. Know which is which before you touch any of this code.