Agent Contracts: A Formal Framework for Resource-Bounded Autonomous AI Systems

TL;DR

Agent Contracts address the lack of formal resource governance for autonomous AI by introducing a rigorous, contract-based framework that unifies input/output, multi-dimensional resource budgets, temporal bounds, and success criteria into a single governance object $C=(I,O,S,R,T,\Phi,\Psi)$. The framework enforces conservation laws for multi-agent delegation, enabling hierarchical, composable coordination with provable budget discipline. The authors formalize contract components, lifecycle, token-budget decomposition, and runtime monitoring, then validate the approach through four experiments showing 90% token savings, 525x reduction in variance, and zero conservation violations in multi-agent delegation, as well as clear quality-resource tradeoffs via contract modes. The results demonstrate that explicit resource governance can make autonomous agent systems predictable, auditable, and safer for production deployment, while signaling future infrastructure needs such as interruptible generation and budget-aware inference to enable hard guarantees.

Abstract

The Contract Net Protocol (1980) introduced coordination through contracts in multi-agent systems. Modern agent protocols standardize connectivity and interoperability; yet, none provide formal, resource governance-normative mechanisms to bound how much agents may consume or how long they may operate. We introduce Agent Contracts, a formal framework that extends the contract metaphor from task allocation to resource-bounded execution. An Agent Contract unifies input/output specifications, multi-dimensional resource constraints, temporal boundaries, and success criteria into a coherent governance mechanism with explicit lifecycle semantics. For multi-agent coordination, we establish conservation laws ensuring delegated budgets respect parent constraints, enabling hierarchical coordination through contract delegation. Empirical validation across four experiments demonstrates 90% token reduction with 525x lower variance in iterative workflows, zero conservation violations in multi-agent delegation, and measurable quality-resource tradeoffs through contract modes. Agent Contracts provide formal foundations for predictable, auditable, and resource-bounded autonomous AI deployment.