Skip to main content

Docker Installation (ddtx)

This guide explains how to set up and use dtx with Docker via the ddtx CLI wrapper.

Using Docker allows you to:

  • Avoid local Python and dependency installations
  • Run dtx fully containerized
  • Easily manage environments and avoid conflicts

Prerequisites​

  • Docker installed and running
  • Python (only for installing dtx wrapper locally)

Verify Docker installation:

docker --version

Ensure Docker daemon is running.

Step 1: Install dtx (for access to ddtx CLI)​

Install the core Python package:

pip install dtx

This installs:

  • dtx — Local CLI (It will not work without troch installed locally)
  • ddtx — Docker CLI wrapper

Step 2: Verify ddtx CLI is available​

Check:

ddtx --help

You should see the same command structure as dtx.

Step 4: Run dtx commands inside Docker​

Now you can run any dtx command inside Docker using ddtx.

Example: Generate a red teaming scope

ddtx redteam scope "My agent description" scope.yml

Example: Generate a red teaming plan

ddtx redteam plan scope.yml plan.yml --dataset STRINGRAY

Example: List available datasets

ddtx datasets list

Notes​

  • No need to install torch or other dependencies locally.

  • ddtx will mount your working directory and .env automatically.

  • Files will be stored in your local filesystem under:

    ~/.dtx/

  • For advanced Docker configuration (volumes, env vars), see the Docker CLI documentation.