roadmap.md

Roadmap

Current Version: 0.4.0

Features

• Core framework with deterministic RNG and replay
• Rock-Paper-Scissors discipline
• Kuhn Poker discipline
• Elo rating system
• Round-robin ladder tournaments
• CLI for running matches and ladders
• JSONL event logging

Future Versions

v0.5.0 - Additional Disciplines

• Tic-Tac-Toe
• Connect Four
• Liar's Dice
• Simplified Texas Hold'em

v0.6.0 - Agent Ecosystem

• Subprocess agent protocol (stdin/stdout)
• Remote agent protocol (HTTP/WebSocket)
• Agent templates for Python, Rust, Go
• LLM agent wrapper

v0.7.0 - Tournament System

• Swiss-system tournaments
• Double elimination brackets
• Persistent rating database
• Tournament scheduling

v0.8.0 - Web Interface

• Match replay viewer
• Live match streaming
• Leaderboard dashboard
• Tournament management UI

v0.9.0 - Advanced Features

• Multi-game ladders
• Agent sandboxing
• Computation budgets
• Move time controls

v1.0.0 - Production Ready

• API stability guarantee
• Performance benchmarks
• Comprehensive documentation
• Docker deployment

Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

Priority Areas

1. New game disciplines
2. Reference agent implementations
3. Documentation improvements
4. Bug fixes and tests

Getting Started

# Clone the repo
git clone https://github.com/wakesync/cmax-arena.git
cd cmax-arena

# Install dependencies
pnpm install

# Build all packages
pnpm build

# Run tests
pnpm test

# Try the CLI
pnpm --filter @cmax/cli start -- run match --game rps --agents random,rps_counter

Design Principles

1. Determinism First - All game logic must be reproducible
2. Simplicity - Prefer simple, understandable implementations
3. Extensibility - Easy to add new games and agents
4. Fairness - No hidden advantages for any participant
5. Transparency - Full audit trail for all matches