Coding Agents with an Obstacle-Aware Harness for Safe Robot Manipulation

Coding Agents with an Obstacle-Aware Harness for Safe Robot Manipulation

• ID: 0d9e270a
• 原文链接: https://arxiv.org/abs/2609.20822
• PDF: https://arxiv.org/pdf/2609.20822v1
• 作者: Bingxin Xu, Yuzhang Shang, Zhen Dong, Emilio Ferrara
• 日期: 2026-09-17
• 更新: 2026-09-17
• 分类: agents
• 来源类型: paper
• 标签: coding-agents, robot-manipulation, safety, harness-design
• 质量评分: 4/5
• 抓取时间: 2026-09-20T04:24:27Z

中文导读

把 coding-agent 范式放到机器人操作安全约束下检验：agent 在多数任务里撞上不许碰的障碍物，轨迹里能推理障碍提示也明令禁止，问题出在规划层安全约束始终变不成优先级 decompose 成 route 与 contact 两阶段定位失败源，并提出 SafeHarnessharness 如何承载安全约束的新证据

为什么值得关注

把 coding-agent 范式放到机器人操作安全约束下检验：agent 在多数任务里撞上不许碰的障碍物，轨迹里能推理障碍提示也明令禁止，问题出在规划层安全约束始终变不成优先级 decompose 成 route 与 contact 两阶段定位失败源...

关键信息

• 论文标题：Coding Agents with an Obstacle-Aware Harness for Safe Robot Manipulation
• 作者：Bingxin Xu, Yuzhang Shang, Zhen Dong, Emilio Ferrara
• arXiv：https://arxiv.org/abs/2609.20822
• 发布时间：2026-09-17
• arXiv 分类：cs.RO, cs.AI, cs.CL, cs.CV
• 关联标签：coding-agents, robot-manipulation, safety, harness-design
• 备注：N/A

English Abstract

Coding agents have emerged as a promising paradigm for robot manipulation: a language model writes the robot controller as a program, and agents built in this way now operate robots without robot-specific training.Whether this paradigm is also safe, however, has not been asked. We evaluate coding agent under a safety constraint, where each task pairs a manipulation goal with an obstacle the robot must not touch. The agent pursues the goal but collides with the obstacle in most cases, treating task completion as its sole objective while neglecting safety. The agent reasons about the obstacle in its traces, and the prompt already forbids touching it, so neither perception nor instruction is at fault; the fault lies in the planning, where the stated constraint never becomes a priority. By decomposing manipulation into a route phase and a contact-rich moment, we locate the source of the failure. Along the route, the model cannot prioritize the safety constraint, having no notion of a clearing route and none of replanning once a chosen route becomes infeasible. At the contact, it is unaware that contact execution is bounded by the same constraint. To close this gap, we present SafeHarness, which equips the model with two obstacle-aware harnesses that enable it to prioritize the safety constraint. Obstacle-aware route planning grounds the objects as bounding boxes and draws candidate routes over them as sequences of waypoints. The agent then plans a route in advance, verifies it, replans when necessary, and only then executes it. Obstacle-aware contact execution instead selects the contact position so that the contact itself avoids the obstacle. SafeHarness attains 71.9% task success and 87.5% collision avoidance, surpassing the previous SOTA by 6.5% and 27.0%, respectively. These results are $2.3\times$ and $1.5\times$ those of the same agent without harnesses.

English Summary

Coding agents have emerged as a promising paradigm for robot manipulation: a language model writes the robot controller as a program, and agents built in this way now operate robots without robot-specific training.Whether this paradigm is also safe, however, has not been asked. We evaluate coding agent under a safety constraint, where each task pairs a manipulation goal with an obstacle the robot must not touch. The agent pursues the goal but collides with the obstacle in most cases, treating task completion as its sole objective while neglecting safety. The agent reasons about the obstacle in its traces, and the prompt already forbids touching it, so neither perception nor instruction is at fault; the fault lies in the planning, where the stated constraint never becomes a priority. By decomposing manipulation into a route phase and a contact-rich moment, we locate the source of the failure. Along the route, the model cannot prioritize the safety constraint, having no notion of a clearing route and none of replanning once a chosen route becomes infeasible. At the contact, it is unaware that contact execution is bounded by the same constraint. To close this gap, we present SafeHarness, which equips the model with two obstacle-aware harnesses that enable it to prioritize the safety constraint. Obstacle-aware route planning grounds the objects as bounding boxes and draws candidate routes over them as sequences of waypoints. The agent then plans a route in advance, verifies it, replans when necessary, and only then executes it. Obstacle-aware contact execution instead selects the contact position so that the contact itself avoids the obstacle. SafeHarness attains 71.9% task success and 87.5% collision avoidance, surpassing the previous SOTA by 6.5% and 27.0%, respectively. These results are $2.3\times$ and $1.5\times$ those of the same agent without harnesses.

Obsidian Notes

• 内容由 opencli arxiv paper 拉取 arXiv 元数据与摘要生成。
• 中文导读、价值判断、关键事实均锚定在条目已有摘要、论文摘要、作者、日期与分类信息上；未补充论文摘要之外的实验细节。
• 抓取时间：2026-09-20T04:24:27Z
• 抓取来源：opencli arxiv paper 2609.20822