From 5ffcd7e28b17dd0e3709cb40650611f1aa3fd281 Mon Sep 17 00:00:00 2001 From: Pengyu Zhang Date: Mon, 10 Aug 2026 19:00:00 -0700 Subject: [PATCH 1/2] docs: add ADR Discovery to the capability overview Adds ADR Discovery as a fifth capability: finding which AI tools are present on an endpoint, as distinct from collecting telemetry from the tools already known to be there. Marked as not included in the current open-source release, consistent with how ADR Prevention is listed. --- README.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index a91e020..840d85e 100644 --- a/README.md +++ b/README.md @@ -6,12 +6,13 @@ ADR is **deployed in production at Uber**, and the accompanying paper was accept ## How ADR secures enterprise AI agents -ADR secures enterprise AI agents through four complementary capabilities: observing agent activity, evaluating defenses, detecting threats, and preventing unsafe actions. +ADR secures enterprise AI agents through five complementary capabilities: observing agent activity, evaluating defenses, detecting threats, preventing unsafe actions, and discovering unsanctioned AI tools. 1. **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. 2. **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. 3. **ADR Detection: Detect risky agent behavior efficiently.** Its two-tier architecture combines high-recall triage with deeper agentic reasoning for suspicious sessions. 4. **ADR Prevention: Stop unsafe actions before they cause harm.** This component is not included in the current open-source release. **Stay tuned.** +5. **ADR Discovery: Find the AI tools present on employee endpoints.** Inventories installed AI applications, CLI agents, local model runtimes, and MCP servers. This component is not included in the current open-source release. **Stay tuned.** ## Repository layout From 84cf1a94e50f1e897b3aa07a418bb05fbd696055 Mon Sep 17 00:00:00 2001 From: Pengyu Zhang Date: Fri, 14 Aug 2026 13:56:09 -0700 Subject: [PATCH 2/2] docs: list ADR Discovery first in the capability overview MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Discovery answers the prior question the other four capabilities assume away — which AI tools are on an endpoint at all — so it reads more naturally ahead of Observability. Reorders the list and the lead-in sentence to match; no wording changes to the items themselves. Co-Authored-By: Claude Opus 5 (1M context) Claude-Session: https://claude.ai/code/session_01GK2DWwdgH8UcGQLMBYpRsT --- README.md | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/README.md b/README.md index 840d85e..129cd76 100644 --- a/README.md +++ b/README.md @@ -6,13 +6,13 @@ ADR is **deployed in production at Uber**, and the accompanying paper was accept ## How ADR secures enterprise AI agents -ADR secures enterprise AI agents through five complementary capabilities: observing agent activity, evaluating defenses, detecting threats, preventing unsafe actions, and discovering unsanctioned AI tools. +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 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. -2. **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. -3. **ADR Detection: Detect risky agent behavior efficiently.** Its two-tier architecture combines high-recall triage with deeper agentic reasoning for suspicious sessions. -4. **ADR Prevention: Stop unsafe actions before they cause harm.** This component is not included in the current open-source release. **Stay tuned.** -5. **ADR Discovery: Find the AI tools present on employee endpoints.** Inventories installed AI applications, CLI agents, local model runtimes, and MCP servers. This component is not included in the current open-source release. **Stay tuned.** +1. **ADR Discovery: Find the AI tools present on employee endpoints.** Inventories installed AI applications, CLI agents, local model runtimes, and MCP servers. This component is not included in the current open-source release. **Stay tuned.** +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, 133 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