How ETF Agent Separates Identity, Authorization, and Resource Isolation
An ETF Agent engineering article explains how a multi-user Agent should distinguish authenticated identity, resource ownership, per-request authorization, targeted sharing, and browser-view isolation. It describes one Run as the authorization lifecycle for background research while emphasizing that the public Open ETF Agent v0.15.2 teaching build is local and single-user rather than a complete multi-user implementation.
The article offers concrete architectural guidance for preventing one user’s research, tools, cached results, or reports from being exposed to another, but its implementation details are self-reported and not independently verified.
一位用户上传了基金资料,让 Agent 在后台分析。另一位用户也登录了同一个应用,并拿到前者报告的链接。服务端应该返回什么?
“已经登录,所以可以访问”显然不够。但把链接藏起来、在页面上过滤列表,也没有解决问题:请求仍然可能直接抵达服务端,旧页面里也可能还留着上一位用户已经加载的内容。
Agent 让这个普通 Web 安全问题变得更长了。研究会跨多次工具调用,用户可能中途退出,后台任务仍在运行;后来打开历史和下载报告,又是新的访问请求。
因此,账户接入的核心不是增加一个登录页,而是让每个读取和效果都说清楚:谁在访问,访问哪份资源,依据哪项授权。 下面用两位用户访问同一项研究的教学场景,解释 ETF Agent 当前的具体分工与取舍。
图示主题:身份、资源授权和浏览器视图隔离。概览按职责组织,不是某次任务的真实调用记录。 完整配图见文末原文链接。