AI Agents Image

Docker image containing multiple AI coding agents and development tools. Tested and used on a Linux host with X11.

This is the image I use daily to work on Afasign, so it includes tools I need that you might not need. If there is interest, I can make it more generally useful, so let me know in the issues. Until then, I'll keep my personal needs as the defaults in this config.

Includes Claude Code, Codex, opencode(+ codenomad and openchamber), kilocode , Antigravity, mistral-vibe, possibly others if they were added after this README's last update.

With some languages and tools: Node.js 24, Rust, Bun, ripgrep, git, asfald (downloader), google chrome

How to use

The file env.sample has a list of environent variables that are used. Copy it to .env and edit it with your keys and information, notably the API keys for the AI agents you want to use.

Build the Docker image

./build.sh

Run

./run.sh /absolute/path/to/your/code

You get a shell in the container in the souce code directory. From there you can start the agent you want.

The container mounts:

• Your code directory at the same location as on your host
• Agents config directories (persisted locally)
• X11/DRI for GUI applications

The container matches host user permissions via USER_NAME, USER_ID, USER_GROUP (defaults to current user).