A scenario to simulate.

A single price path for an asset.

Full simulation output.

Trains ATLAS agents using MiroFish future scenarios.

A complete future scenario with multiple asset paths.

Generates synthetic future scenarios from swarm predictions.

Meta-weighting layer that blends recommendations from multiple agent cohorts based on their rolling accuracy.

Claude-based multi-agent simulation. Simulates market agent interactions without external dependencies.

Main interface for MiroFish integration with ATLAS. Usage: bridge = MiroFishBridge() # Generate and run simulation

Scores past predictions against actual outcomes. Tracks accuracy by scenario type for autoresearch integration.

Generates simulation scenarios from current market context and agent debates.

Collects current market data and formats as seed material for simulation.

A prediction from simulation.

A tail risk identified by simulation.