ADR: Agentic AI Detection and Response

ADR智能体AI检测与响应是一个面向AI智能体的企业安全系统。它帮助组织保护面向员工的智能体如Cursor、Claude Code和Codex以及面向客户的智能体如AI客服助手。ADR (Agentic AI Detection and Response) is an enterprise security system for AI agents. It helps organizations secure employee-facing agents such as Cursor, Claude Code, and Codex, as well as customer-facing agents such as AI support agents.已经在 Uber 部署了GitHub - uber/ADR: ADR secures enterprise AI agents through observability, security benchmarking, and threat detection. Deployed at Uber. · GitHub论文地址[2605.17380] ADR: An Agentic Detection System for Enterprise Agentic AI SecurityHow ADR secures enterprise AI agentsADR 通过四项互补能力来保障企业级 AI 智能体的安全观察智能体活动、评估防御措施、检测威胁以及阻止不安全行为。ADR 可观测性了解 AI 智能体正在做什么以及为什么这样做。在生产环境中ADR 能够捕获智能体的意图、工具使用情况和执行轨迹覆盖 macOS、Linux 和 Windows 上的 7 种以上 AI 编程工具以及内部自动化系统和面向客户的客服智能体。ADR 基准测试在真实的企业环境下测试智能体安全性。ADR-Bench 包含 300 多个任务、133 个 MCP 服务器并覆盖所有 17 种智能体攻击技术。ADR 检测高效检测有风险的智能体行为。其双层架构结合了高召回率的初步筛选和对可疑会话的深度智能体推理。ADR 防护在造成危害之前阻止不安全行为。此组件未包含在当前的开源版本中。敬请期待。ADR secures enterprise AI agents through four complementary capabilities: observing agent activity, evaluating defenses, detecting threats, and preventing unsafe actions.ADR Observability: Understand what AI agents are doing and why.In production, ADR captures agent intent, tool use, and execution traces across 7 AI coding tools on macOS, Linux, and Windows, as well as internal automation and customer-facing support agents.ADR Benchmark: Test agent security under realistic enterprise conditions.ADR-Bench includes 300 tasks, 133 MCP servers, and coverage of all 17 agent attack techniques.ADR Detection: Detect risky agent behavior efficiently.Its two-tier architecture combines high-recall triage with deeper agentic reasoning for suspicious sessions.ADR Prevention: Stop unsafe actions before they cause harm.This component is not included in the current open-source release.Stay tuned.Repository layoutThis repository contains the open-sourceADR Sensor,ADR-Bench, andADR Detectordescribed in the paper. The offlineADR Explorerengine, which hardens ADR Detection through pre-deployment red teaming, is not included here.PathADR componentDescriptionSensor/ADR ObservabilityCollect and normalize agent telemetry from Claude Code, Cursor, Codex, opencode, Claude Desktop, and othersDetection/ADR Benchmark DetectionDual-agent detector, 133 MCP servers, 303 benchmark tasks, baselines, figure scriptsdocs/REPRODUCIBILITY.mdEvaluationStep-by-step workflow to reproduce benchmark detection and paper figures---