A proof-of-concept implementation of an AI-driven cybersecurity wargame on the Qubic L1 blockchain using uPoW jobs.

This project implements attack and defense AI agents that compete in a sandboxed environment, with results recorded on-chain. The system ensures deterministic execution, complete network isolation, and reproducible results.

# Setup environment
./scripts/setup.sh

# Run a test round
./scripts/run_round.sh

# Verify determinism
./scripts/assert_determinism.sh

# Check network isolation
./scripts/assert_no_egress.sh

• Smart Contracts: Job registration and result verification (C++)
• Relay Service: Worker-to-blockchain communication (Go)
• Workers: Attack and Defense AI implementations
• Sandbox: Isolated Docker/KVM environments
• Storage: MinIO for off-chain artifacts
• Dashboard: Web UI for monitoring

# Unit tests
make test-unit

# Integration tests
make test-integration

# End-to-end tests
make test-e2e

# All tests
make test

See /docs for detailed documentation:

• API Specification
• Threat Model
• Architecture

SPDX-License-Identifier: MIT