DarwinX：通过自然选择进化智能体调控框架

摘要

智能体（LLM agent）的能力不仅取决于模型权重，还取决于其运行框架（harness）：提示、工具、技能和控制流。自我改进循环已经能够编辑运行框架，但单谱系搜索具有路径依赖性，且局部改进往往会导致其他任务回退。我们提出DarwinX，它将自我进化视为在冻结模型条件下对运行框架种群的筛选：一个“保留并扩展”契约只允许那些在不导致回退的前提下扩展覆盖范围的变体；一个档案库保存替代谱系以供重组；来自失败、教师和自身的证据共享同一个编辑接口。适应性来自每个基准自身的验证器：不需要黄金解答，也不需要人工挑选的胜者。在四个逐步将进化信号与测试分离的基准上，单个循环平均增加约17个百分点：Terminal-Bench 2.1在匹配基座上提升7.7个百分点至83.2%，在更强基座上达到84.7%的已验证前沿；TerminalWorld的保留划分达到68.3%，超过所有现成智能体；WebArena-Infinity真实任务pass@1经审计干净后从43.5%提升至93.0%；一个Terminal-Bench 2.1运行框架原样迁移至SWE-bench Verified仍然有效。真正进化的是通用智能体能力，而非针对特定基准的补丁，因此它能够经受任务、验证器和基础模型的改变。冻结的模型未必是固定的智能体：运行框架选择将评估计算转化为持久能力。

English

An LLM agent's capability depends not only on model weights but on its harness: prompts, tools, skills, and control flow. Self-improvement loops already edit harnesses, yet single-lineage search is path-dependent and local wins often regress other tasks. We introduce DarwinX, which treats self-evolution as selection over a population of harnesses with the model frozen: a preserve-and-extend contract admits only variants that extend coverage without regressing, an archive keeps alternative lineages for recombination, and failure-, teacher-, and self-derived evidence share one edit interface. Fitness comes from each benchmark's own verifier: no gold solutions, no hand-picked winners. Across four benchmarks that progressively separate the evolution signal from the test, one loop adds about 17 points on average: Terminal-Bench 2.1 rises +7.7 to 83.2% on a matched base and to the verified frontier at 84.7% on a stronger one; TerminalWorld's held-out split reaches 68.3%, ahead of every off-the-shelf agent; WebArena-Infinity real-task pass@1 rises from 43.5% to 93.0% audit-clean; and a Terminal-Bench 2.1 harness transfers unchanged to SWE-bench Verified. What evolves is general agent competence, not benchmark-specific patches, so it survives changes of task, verifier, and base model. A frozen model need not be a fixed agent: harness selection turns evaluation compute into durable capability.