AECP：面向多智能体代码生成的工件专属通信协议

中文摘要 AI 辅助

随着AI智能体日益应对复杂的仓库级编码任务，将工作分配给多个智能体是超越单一智能体能力的自然扩展方式。为协调相互依赖的工作，这些智能体共享发现结果并就模块间的接口达成一致。然而，交换的信息往往仅作为上下文，留给各个智能体自行解读并将其纳入后续工作。因此，共享的发现可能未被使用，偏离接口协议的情况可能未被察觉，从而削弱协作的可靠性和效率。这促使将部分协调责任从个体智能体转移到执行框架（harness）。为使共享信息在执行过程中可操作，我们引入了工件专属通信协议（Artifact-Exclusive Communication Protocol, AECP）。AECP要求智能体仅通过结构化工件进行通信，并规定了框架如何处理这些工件。框架在智能体访问相关代码时提供发现结果，筛查实现与记录的接口承诺之间的不匹配，并要求受影响的智能体重访修订后的协议。这些协调步骤成为框架执行的一部分，而非智能体必须根据先前消息主动采取的行动。在Doc2Repo、NL2Repo和CodeProjectEval上，使用包括Opus-4.8和DeepSeek-V4-Flash在内的闭源和开源模型，与使用自由形式智能体间消息的智能体团队相比，AECP将平均测试通过率提高了28.2%，并将平均墙钟时间（wall time）减少了16.5%。工件专属通信还阻止了恶意指令在智能体之间的中继，将其到达其他智能体的频率从95%降至0%，以及这些智能体执行这些指令的频率从40%降至0%。

英文摘要

As AI agents increasingly tackle complex repository-level coding tasks, distributing work across multiple agents is a natural way to scale beyond the capabilities of a single agent. To coordinate their interdependent work, these agents share findings and agree on interfaces between modules. However, exchanged information often serves only as context, leaving individual agents to interpret it and incorporate it into subsequent work. Consequently, shared findings may go unused and deviations from interface agreements may go undetected, undermining the reliability and efficiency of collaboration. This motivates moving part of the coordination responsibility from individual agents to the execution harness. To make shared information actionable during execution, we introduce the Artifact-Exclusive Communication Protocol (AECP). AECP requires agents to communicate exclusively through structured artifacts and specifies how the harness processes them. The harness supplies findings when agents access relevant code, screens implementations for mismatches with recorded interface commitments, and requires affected agents to revisit revised agreements. These coordination steps become part of harness execution rather than actions that agents must initiate from prior messages. Across Doc2Repo, NL2Repo, and CodeProjectEval, using closed- and open-source models including Opus-4.8 and DeepSeek-V4-Flash, AECP improves average test pass rate by 28.2% and reduces average wall time by 16.5% relative to an agent team using free-form inter-agent messages. Artifact-exclusive communication also blocks the relay of malicious instructions between agents, reducing how often they reach other agents from 95% to 0% and how often those agents act on them from 40% to 0%.

发表机构

• University of Central Florida（中佛罗里达大学）
• AWS AI Labs（AWS AI实验室）

机构由 AI 辅助整理，请以论文原文为准。