https://guijiagi.com/tags/%E6%9D%83%E9%99%90%E6%8E%A7%E5%88%B6/Recent content in 权限控制 on 硅基 AGI · 智能体学习与测评Hugozh-cn本站内容采用 CC BY-NC-SA 4.0 国际许可协议授权Thu, 16 Jul 2026 11:34:00 +0800https://guijiagi.com/posts/b1-0e81ff69/Thu, 16 Jul 2026 11:34:00 +0800https://guijiagi.com/posts/b1-0e81ff69/探讨构建可信赖AI Agent的安全架构设计，涵盖权限控制、行为审计、风险分级与人类监督机制https://guijiagi.com/posts/article-46/Mon, 13 Jul 2026 00:30:00 +0800https://guijiagi.com/posts/article-46/从进程隔离到能力管控，AI Agent安全沙箱的架构设计与实现https://guijiagi.com/posts/b2-91087680/Mon, 13 Jul 2026 00:30:00 +0800https://guijiagi.com/posts/b2-91087680/从进程隔离到能力管控，AI Agent安全沙箱的架构设计与实现https://guijiagi.com/posts/agent-security-audit-jailbreak-to-permission/Tue, 30 Jun 2026 09:30:00 +0800https://guijiagi.com/posts/agent-security-audit-jailbreak-to-permission/全面解析Agent系统的安全威胁模型与防护方案，涵盖越狱攻击、提示注入、权限逃逸、数据泄漏等核心安全问题