How it works
Portable Docker sandboxes that let coding agents run unrestricted, safely. The full technical description of Agent Lab — what it does, what it is made of, and what state it is in.
The premise
Agent Lab lets Codex and Claude Code run in their fully unrestricted modes while Docker confines them to the one project folder you explicitly mount. The dangerous setting becomes the safe setting because the blast radius is the container.
Logins for the agent CLIs, GitHub and Google Cloud live in Docker-only named volumes, so credentials survive between sessions without ever sitting in the sandbox filesystem. One install script sets up a new machine.
What is included
Unrestricted agent modes confined to an explicitly mounted project directory
Credentials isolated in Docker-only named volumes, never in the sandbox tree
Single install script; images build on first use
One command surface across both agent CLIs for logins and MCP servers
How it is built
Agent Lab is built on Docker, Shell, PowerShell, Claude Code and Codex.
The repository is 44% Shell, 27% JavaScript, 24% PowerShell, 5% Dockerfile by volume — a description of what the work actually consists of rather than what it is filed under.
It carries 37 commits since Aug 2026, most recently changed in Aug 2026.
The method behind it
Every project here starts as a written specification, and the specification stays authoritative once code exists. The repository keeps a build plan with live status, an append-only decision log recording where the implementation departed from the plan, and a handoff document describing the true current state — including the parts that are not finished.
That is why a portfolio this wide is not a shallow one, and it is why this page can describe the product without overstating it.
Current state
In development. Under active work right now.
The source is public. Read it rather than taking this page for it.
Next
Put Agent Lab in your pipeline
Ask about integrating it with the repositories and agents you already run.