AI agents and trading strategies
Multi-agent research powered by LangGraph + Claude Sonnet 4.6. Supports Flash, Standard, Pro, and Ultra research modes.
Reviews portfolio performance, summarizes P&L, flags anomalies. Feeds the morning digest.
Scans GitHub for security advisories, failed CI, and dependency drift across active projects.
Drafts TikTok/Instagram content for WooVision. All output requires approval before posting.
Buys mispriced directional legs where Kalshi threshold market prices violate monotonicity. Stop-loss 20%, take-profit 40%.
Binary arbitrage on Kalshi bracket groups where YES prices sum > 1.02. Buys NO, holds to resolution.
Cross-platform arbitrage using KL-divergence to detect distributional mispricing between Kalshi and Polymarket.
Replicates trades from skilled market participants.
Uses Claude to estimate outcome probabilities, trades when confidence exceeds market price.
Statistical mean reversion strategy.
Detects informed trading flows and trades ahead of market moves.
Fades longshot positions by selling overpriced low-probability outcomes.
Market resolution tracking and settlement utility. Non-trading support tool.
Weather market arbitrage using Open-Meteo forecasts. Currently disabled.
High-frequency micro-arbitrage on tiny mispricings. Stale/inactive.
Monotonicity violation arbitrage. Superseded by mono-arb.