智能体
developer.nvidia.com ·
@Elizabeth Goodman
NVIDIA OpenShell 参考架构：规模化治理 AI Agent 工作负载
文章讨论自主 Agent 长时间执行复杂任务时访问文件、工具、互联网并派生其他 Agent，导致风险累积：可能下载代码、构建新工具、申请额外权限并触及更多基础设施，且行动路径在执行中动态确定。NVIDIA 提出 OpenShell 参考架构以安全与治理此类 Agent 工作负载，但摘要未展开具体机制。
AI推荐理由 聚焦长时程自主 Agent 的复合风险（下载代码、自建工具、申请权限、扩大基础设施接触面），提出 OpenShell 参考架构用于规模化安全与治理，问题定义具体，对 Agent 平台建设者有实用价值。局限是正文片段仅给出风险描述，未提供架构组件、治理策略或部署案例细节。
70
正文
Autonomous agents can work for hours on complex tasks, accessing files, tools, the internet, and spawning other agents to collaborate with. Executing complex...
Autonomous agents can work for hours on complex tasks, accessing files, tools, the internet, and spawning other agents to collaborate with. Executing complex tasks over long time horizons leads to compounding risk: the agents may download code, build new tools, request additional privileges, and touch more of the underlying infrastructure. The agents’ action paths are determined as they work…
Source