共同演化之代理框架與模型：同策略修正助較弱模型在模仿失敗之處迎頭趕上

摘要

代理框架（圍繞模型的系統提示、工具集、執行鉤子與上下文管理鷹架）是代理任務成功的關鍵決定因素。自動化代理框架演化能讓較小模型以極低於前沿模型的成本，在領域特定任務上表現良好。由於代理框架與模型權重皆會形塑行為，我們探討應如何結合代理框架演化與輕量微調。在七項企業代理任務中，我們首先以較弱模型演化出代理框架，接著發現較強專家往往能更有效地使用該框架，這顯示專家監督可能彌補剩餘差距。然而，在演化後的框架下，以專家的完整軌跡訓練較弱模型卻適得其反：在 Qwen3-Coder 與 Gemma 4 上，所有七項任務的效能皆退步 4 至 30 分，儘管相同程序在未演化框架下有所幫助。我們的分析顯示，模仿雖能轉移知識並增加鷹架使用，卻破壞了模型與框架的契合度：較弱模型採用了專家的規劃策略，卻不具備執行該策略的能力，且不再匹配圍繞其原生規劃風格所演化出的框架。因此，我們開發了一套同策略專家修正流程，由元層級 MLE 代理自動化執行，該流程能定位較弱模型自身 rollout 中的失敗回合，並要求專家僅改寫該回合。此舉保留了模型的規劃風格，並結合了代理框架演化與模型適應的增益。我們的研究結果識別並解決了框架與權重更新之間的衝突來源，為領域特定企業任務上的經濟共同演化提供了一套保持相容性的方法。

English

Agent harnesses (the system prompt, tool set, execution hooks, and context-management scaffolding around a model) are a critical determinant of agentic task success. Automated harness evolution can enable smaller models to perform well on domain-specific tasks at a fraction of frontier-model cost. Since both the harness and model weights shape behavior, we ask how harness evolution and lightweight fine-tuning should be combined. Across seven enterprise agent tasks, we first evolve a harness with the weaker model, then find that a stronger expert often uses it more effectively, suggesting expert supervision could close the remaining gap. However, training the weaker model on the expert's complete trajectories under the evolved harness backfires: performance regresses on all seven tasks by 4 to 30 points across Qwen3-Coder and Gemma 4, even though the same procedure helps under the unevolved harness. Our analysis shows that imitation transfers knowledge and increases scaffold usage, but disrupts model-harness fit: the weaker model adopts the expert's planning strategy without the competence to execute it and no longer matches the harness evolved around its native planning style. We therefore develop an on-policy expert-correction pipeline, automated by a meta-level MLE agent, that localizes the failing turn in the weaker model's own rollout and asks the expert to rewrite only that turn. This preserves the model's planning style and combines the gains of harness evolution and model adaptation. Our results identify and resolve a source of contention between harness and weight updates, yielding a compatibility-preserving recipe for economical co-evolution on domain-specific enterprise tasks.