Agent Lab: A Queue UI for AI Work

This build lab focuses on the queue-based workbench that lets students submit AI jobs and watch them move through states. The goal is not to copy a private machine setup. The goal is to learn the architecture pattern well enough to build a small, classroom-safe version.

A queue UI separates job creation from job execution: users submit structured work, workers process it, and the interface shows status without blocking the browser.

Build the small version

1. 1Draw or write a queue state machine with draft, queued, running, needs-review, done, and failed states.
2. 2Mark which parts are user-facing, which parts are internal, and which parts require approval.
3. 3Choose one low-risk workflow and implement only that workflow first.
4. 4Add one failure case before adding a second feature.
5. 5Write a short operator note: what the agent may do, what it must ask about, and what it must never do.

job_states: draft -> queued -> running -> needs_review -> done queued -> failed running -> failed job fields: id, owner, prompt, provider, workspace, status, result, error, created_at, updated_at

The big idea: queue is not decoration. It is part of the product architecture students need before an agent becomes safe enough to use with real people.