📢 Notice: This repository is a Component Submodule of the Vindicta Platform Monorepo. For local development, testing, and dependency resolution, please clone the root vindicta-platform repository.

Predictive Models & Strategic Meta Analysis.

The Oracle Domain provides the AI-driven foresight of the Vindicta Platform. It evaluates army lists, predicts match outcomes, and resolves complex strategic debates through a multi-agent consensus system.

• Debate Engine: Multi-agent system where different tactical personas (Tech-Priest, Chaos Sorcerer, etc.) debate game mechanics.
• List Grader: Automated evaluation of competitive army lists against current meta-trends.
• Outcome Predictor: Statistical models for win probability and tactical risk assessment.

This repository is managed with uv.

# Install dependencies
uv sync

# Run the debate engine demo
uv run src/vindicta_oracle/demo.py

• src/vindicta_oracle/agents/: Individual strategic personas and reasoning logic.
• src/vindicta_oracle/engine.py: The consensus and orchestration logic for debates.
• src/vindicta_oracle/grader.py: Army list evaluation logic.

Part of the Vindicta Platform. See vindicta-foundation for global standards.