Harness-S：借助兄弟 Harness 训练更通用的智能体
Harness-S: Learning from Sibling Harnesses for Generalizable Agent Training
一篇训练智能体的新论文，思路挺巧：基座 harness 跑挂了就让一堆兄弟 harness 救场，再把本事转移回策略本身，AppWorld 等三个基准都涨了。
论文提出 Harness-S 框架，通过多个兄弟 harness 修复基础执行系统的失败案例，并把这种能力迁移进单一策略。方法包含三个组件：Directional Harness Evolution 构建固定的 Sibling Harness Pool，Complementary Harness Rescue Training 仅在基座 harness 初始尝试全部失败时调用辅助 harness 生成新轨迹，Adaptive Harness Router 用多 LLM 投票选出 Top-k 变体。在 AppWorld、SWE-Gym 和 DBBench 三个基准上，训练后的策略回到原始 harness 下仍显著超过仅用基座 harness 的策略训练。将 Router 复用于 Rank-1 推理还能在不增加执行 rollout 数量的情况下带来额外收益。
Harness-S: Learning from Sibling Harnesses for Generalizable Agent Training
Harness evolution from a common base exhibits Locality, Complementarity, and Nonstationarity: variants repair different failures, and their utility changes as the policy learns. Rather than seeking a universally superior harness, we introduce Harness-S to selectively rescue failures of a base execution system through sibling harnesses and transfer the resulting capabilities into a single policy that remains effective under the original harness. Directional Harness Evolution develops specialized sibling harnesses along distinct repair directions, forming a reusable Sibling Harness Pool that remains fixed during policy training. Complementary Harness Rescue Training invokes this pool only when all initial attempts under the base harness fail. The same policy generates fresh rollouts under selected auxiliary harnesses, and these trajectories replace the failed group for reinforcement learning. An Adaptive Harness Router selects the Top-k variants using Harness Behavior Profiles and voting across multiple LLMs. Policy-aware profile re-anchoring refreshes this evidence as the policy learns. Across AppWorld, SWE-Gym, and DBBench, the trained policy substantially outperforms policy-only training when returned to the original base harness. Behavioral analyses and held-out cases support the transfer of harness-elicited behaviors into the policy itself. Reusing the router for Rank-1 inference provides additional gains at the same executor rollout count.