携手 NVIDIA，让企业更好地掌控 AI Agent
英文标题：Giving companies more control over their AI agents, with NVIDIA
NVIDIA 发布开放式 Agent 安全平台 Open Agent Safety Platform，Anthropic 与 NVIDIA 合作，为 AI Agent 技术栈增加安全与控制层。Claude Managed Agents 将凭据存放在独立保险库中，Agent 无法直接查看；NVIDIA OpenShell 则通过运行时策略限制 Agent 可执行的操作、可访问的文件、网络和数据。两者结合后，企业可限制 Agent 权限、审查其行为日志，并验证相关限制是否真正生效。该方案面向正在将 AI 从问答工具扩展为可访问专有数据、跨业务部门执行复杂任务的企业级 Agent 应用。
22.7 KB
403 B
49.2 KB
13.8 KB
17.7 KB
10.2 KB
49.1 KB
中文
NVIDIA 今日宣布推出 Open Agent Safety Platform（开放式 Agent 安全平台），这是一个用于加强 AI 安全的开放软件平台和参考系统设计。Anthropic 已与 NVIDIA 合作，为 Agent 技术栈带来额外的安全与控制层。
Claude Managed Agents 是一套可组合的 API，用于大规模构建和部署生产级 Agent。它会将 Agent 所需的凭据保存在保险库中，因此 Agent 永远不会看到这些凭据。开源的 NVIDIA OpenShell 软件旨在控制 Agent 在工作期间可以执行和访问的内容。使用 Managed Agents 和 OpenShell 的客户可以限制 Agent 的行为、审查 Agent 已完成的操作，并确认这些限制已生效。
企业正从使用 AI 回答问题，转向部署能够跨业务部门处理复杂工作、使用专有数据并代表用户采取行动的 Agent。随着模型能力提升，Agent 的应用场景不断扩大，获得的访问权限也越来越多。Agent 获得的访问权限越多，其所在企业就越需要控制和核查它的行为。
分层防护
防护始于模型内部的安全措施。Managed Agents 和 NVIDIA Open Shell 增加了位于模型外部的限制措施，用于约束 Agent 的实际行为。每一层均被设计为能够独立执行自身限制，因此整体防护不依赖于任何单一层。各层均采用模块化设计，企业可选择适合其部署环境的组件。
Claude Managed Agents 负责执行工作并保管凭据
使用 Managed Agents 时，Agent 循环运行在独立于沙箱的服务器上；沙箱则是实际执行工作的隔离环境。凭据，即密码和访问密钥，被保存在单独的保险库中，因此 Agent 永远不会看到它们。
Managed Agents 还提供审计追踪功能，用于记录每个 Agent 执行过的操作，并可与企业现有的访问控制系统集成。企业可以自带沙箱配置，并自行选择沙箱的运行位置和运行方式。
NVIDIA OpenShell 规定 Agent 可访问的范围
OpenShell 是 NVIDIA 推出的开源安全运行时软件。它负责治理和监控所有 AI Agent 行为，并对每一项操作执行策略。除非规则明确允许，否则 OpenShell 会阻止一切操作。它会检查 Agent 尝试使用的每一项工具，并将规则应用于 Agent 访问的文件、网络连接和数据。相关规则在 Agent 外部执行，OpenShell 会记录其允许或阻止的每一个决策。
团队可以从较为严格的权限范围开始，审查日志，并使用 Claude 持续收紧规则，使权限逐步接近完成任务所需的最小访问范围。随后，OpenShell 的策略证明器会利用数学证明，确认 Agent 在团队所编写规则下能够访问的范围。
Claude Managed Agents 包含的功能
- 生产级 Agent：由平台处理安全沙箱、身份验证和工具执行。
- 可自主运行数小时的长时会话；即使发生断开连接，进度和输出仍会保留。
- 多 Agent 编排：Agent 可以启动并指挥其他 Agent，以并行处理复杂工作。
- 可信治理：通过内置的范围限定权限、身份管理和执行追踪，让 Agent 能够访问真实系统。
团队如何使用 Managed Agents
- Notion 让团队可以在其工作空间内将工作交给 Claude。工程师使用它交付代码，其他员工则使用它制作网站和演示文稿。在团队共同处理结果的同时，数十项任务可以并行运行。
- Rakuten 在工程、产品、销售、市场营销和财务等部门运行专业 Agent，每个 Agent 都能在一周内完成部署。
- Asana 构建了 AI Teammates，即与 Asana 项目中的人员协同工作的 Agent，能够承担任务并起草交付成果。借助 Managed Agents，该团队得以比原本更快地加入高级功能。
可用性
Managed Agents 即日起可用。它可以运行在由您控制的沙箱中，该沙箱既可部署在您自己的基础设施上，也可通过托管服务提供商运行。NVIDIA OpenShell 以 Apache 2.0 许可证开源，可通过 GitHub 和 NVIDIA 的开发者资源页面获取。
English / 原文
NVIDIA today announced the Open Agent Safety Platform, an open software platform and reference system design for strengthening AI security. Anthropic has collaborated with NVIDIA to bring additional layers of security and control to the agent stack.
Claude Managed Agents, a suite of composable APIs for building and deploying production-grade agents at scale, holds the credentials an agent needs in a vault so the agent never sees them. Open source NVIDIA OpenShell software is designed to control what the agent can execute and reach while it works. Customers who are using Managed Agents with OpenShell can limit what an agent can do, review what the agent did and confirm that the limits are in place.
Companies are moving from using AI to answer questions to deploying agents that handle complex work across business units, use proprietary data and take actions on behalf of users. As models improve, agents find more uses and get more access. The more access an agent has, the more its company needs to control and check what it does.
Protection in layers
Protection starts with safeguards inside the model. Managed Agents and NVIDIA Open Shell add limits that sit outside the model and apply to what the agent does. Each layer is designed to enforce its limits independently, so protection doesn't depend on any single layer. The layers are modular, so companies can adopt the ones that fit their setup.
Claude Managed Agents does the work and holds the credentials
With Managed Agents, the agent loop runs on a separate server from the sandbox, the isolated environment where the work happens. Credentials, meaning passwords and access keys, are held in a separate vault, so the agent never sees them.
Managed Agents also provides audit trails, which record what each agent did, and integration with a company's existing access controls. Companies can bring their own sandbox setup and choose where and how it runs.
NVIDIA OpenShell sets what an agent can reach
OpenShell is open source secure runtime software from NVIDIA. It governs and monitors all AI agent behavior and enforces policies for every action. OpenShell blocks everything unless a rule allows it. It checks each tool an agent tries to use and applies rules to the files, network connections and data the agent accesses. The rules are enforced outside the agent, and OpenShell logs every decision it allows or blocks.
Teams can start with narrow permissions, review the log, and use Claude to tighten the rules toward the least access a task needs. OpenShell's policy prover then uses mathematical proof to confirm what the agent can reach under the rules the team wrote.
What Claude Managed Agents includes
- Production-grade agents with secure sandboxing, authentication, and tool execution handled for you.
- Long-running sessions that operate autonomously for hours, with progress and outputs that persist even through disconnections.
- Multi-agent orchestration, where agents can spin up and direct other agents to parallelize complex work.
- Trusted governance, giving agents access to real systems with scoped permissions, identity management, and execution tracing built in.
How teams use Managed Agents
- Notion lets teams hand work to Claude inside their workspace. Engineers use it to ship code, and other employees use it to produce websites and presentations. Dozens of tasks can run in parallel while the team works on the results together.
- Rakuten runs specialist agents across engineering, product, sales, marketing, and finance, each deployed within a week.
- Asana built AI Teammates, agents that work alongside people in Asana projects, take on tasks and draft deliverables. Using Managed Agents, the team added advanced features faster than it could have otherwise.
Availability
Managed Agents is available today. It can operate in a sandbox you control either running on your own infrastructure, or with a managed provider. NVIDIA OpenShell is open source under the Apache 2.0 license and available on GitHub and NVIDIA's developer resources page.