ADR: Agentic AI Detection and Response

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, Codex, GitHub Copilot CLI, and DeepSeek Harness, as well as customer-facing agents such as AI support agents.

ADR is deployed in production at Uber, and the accompanying paper was accepted to MLSys 2026: Paper PDF · Slides PDF

How ADR secures enterprise AI agents

ADR secures enterprise AI agents through five complementary capabilities: discovering unsanctioned AI tools, observing agent activity, evaluating defenses, detecting threats, and preventing unsafe actions.

1. ADR Discovery: Find the AI tools present on employee endpoints. Inventories installed AI applications, CLI agents, IDE extensions, local model runtimes, and MCP servers, and flags unknown surfaces for review.
2. 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.
3. ADR Benchmark: Test agent security under realistic enterprise conditions. ADR-Bench includes 300+ tasks, 134 MCP servers, and coverage of all 17 agent attack techniques.
4. ADR Detection: Detect risky agent behavior efficiently. Its two-tier architecture combines high-recall triage with deeper agentic reasoning for suspicious sessions.
5. ADR Prevention: Stop unsafe actions before they cause harm. This component is not included in the current open-source release. Stay tuned.

Repository layout

This repository contains the open-source ADR Discovery, ADR Sensor, ADR-Bench, and ADR Detector described in the paper. The offline ADR Explorer engine, which hardens ADR Detection through pre-deployment red teaming, is not included here.

Quick start: ADR Detection

git clone https://github.com/uber/ADR
cd ADR/Detection
uv sync
export ANTHROPIC_API_KEY="..." OPENAI_API_KEY="..."

Default detector is adr (ADR dual-agent). For keyless smoke tests, use --detector llamafirewall (see Detection/README.md).

See docs/REPRODUCIBILITY.md for the full evaluation workflow (inflate packed benchmark → run detectors → plot figures).

Component documentation:

ADR Sensor also captures Gemini CLI session journals on macOS, Linux, and Windows, including tool results and nested subagent sessions.

• Discovery/README.md: endpoint inventory, probes, and the fingerprint catalog
• Sensor/README.md: telemetry collection and unified schema
• Detection/README.md: ADR-Bench, detector baselines, MCP infrastructure

Citation

@inproceedings{li2026adr,
  title={ADR: An Agentic Detection System for Enterprise Agentic AI Security},
  author={Li, Chenning and Hu, Pan and Xu, Justin and Ozbas, Baris and Liu, Olivia and Van, Caroline and Li, Manxue and Zhou, Wei and Alizadeh, Mohammad and Zhang, Pengyu and Sriramadhesikan, KK and Zhang, Ming},
  booktitle={Proceedings of the Ninth Conference on Machine Learning and Systems},
  year={2026}
}

Or use CITATION.cff.

Star History

License

Apache License 2.0. See LICENSE. Detection/benchmark/agentdojo/ is vendored third-party code under its own LICENSE (MIT).

Data notice

Detection/ includes synthetic benchmark fixtures (fake credentials, emulated environments, prompt-injection scenarios) for defensive security research only. Details: docs/OPEN_SOURCE_REVIEW.md.