Skip to content
XORCISE.AI
Backend running

Get started with XORCISE.AI

Evaluate how well cyber AI agents perform on real cybersecurity missions — declare an agent, pull a mission, and watch it work end to end.

1Register an agentNext · Point XORCISE at your harness.2Install a missionPull from the library or ingest a bundle.3Start a runWatch it work in real time.4Review the resultScored assertions + narrative.

Finish setup

Complete the required steps below to run your first evaluation.

  • XORCISE.CORE reachable

    Checking

    Checking…

  • Database ready

    Checking

    Checking…

  • Docker available

    Checking

    Checking…

  • An agent is registered

    Checking

    Checking…

  • A mission is available

    Checking

    Checking…

Recommended

  • Judge model configured

    Not configured

    Deterministic evaluation is available, but model-based judging is disabled.

    Configure in Settings

  • Remote catalog connected

    Not configured

    Connect it to browse and pull missions.

    Configure in Settings

Quick start

Start a Run

Create your first evaluation.

Browse Missions

Explore available targets.

Read Documentation

Setup, traces, results.