Zum Hauptinhalt springen

AgentRidge Documentation

A desktop application that runs authorized penetration tests on its own — an autonomous brain agent that plans, delegates, and reports, with a fleet of specialized sub-agents doing the work. This is the complete guide for the macOS, Windows, and Linux desktop app.

macOS 12+Windows 10 / 11Linux · Ubuntu 22.04+Local-first runtime

What is AgentRidge?

AgentRidge is desktop Mission Control for autonomous penetration testing, built by Azuro Software. You describe an engagement — a client, a target, a scope, and the kinds of tests you want — and AgentRidge carries it out end to end: reconnaissance, enumeration, controlled exploitation, evidence collection, and a client-ready report.

Under the hood it is an agentic system. A single orchestrating agent — the Agent Ridge Principal — reasons about the engagement and coordinates specialized Ridge Sub-Agents that generate commands, parse tool output, and turn raw results into confirmed findings. Heavy reasoning runs locally on your machine through the Ridge inference engine, so sensitive engagement data stays on the operator's workstation.

White-label

AgentRidge is a white-label product. Throughout the app and this documentation the components are named Agent Ridge Principal, Ridge Sub-Agents, and the Ridge inference engine — never any third-party brand.

What makes it different

  • Truly autonomous. The Principal doesn't just suggest commands — it runs the full attack or audit path and pushes each lead further until the engagement's definition of done is met.
  • Two engagement personalities. Run in Auditor mode for proportionate, compliance-oriented checks, or Attacker mode for maximum-depth adversary simulation.
  • Unrestricted security reasoning. The sub-agents operate without the refusal behavior that normally blocks offensive-security work, so they engage with real exploitation tasks under your written mandate. See Autonomy & agent unlocking.
  • Local-first. The reasoning engine and runtime live on your machine. (Local-first design.)
  • Report-ready. Every confirmed finding becomes structured data that feeds live cartography, a vulnerability register, and an exportable PDF report.

Start here

Authorized use only

AgentRidge is a tool for authorized security testing. Every mission requires a confirmed written mandate before it can launch, and the app is built for use against systems you own or have explicit permission to test. Read Authorized use before your first engagement.