Learning Lab
Codingadvanced 25 min

Rehearse coding work safely

Coding agents become powerful when they can read files and run commands. This lesson teaches the trust instruments that should become visible before real-machine access is enabled.

0 of 3 observations0%

Choose an isolated workspace

Use a disposable or version-controlled project. Confirm the displayed workspace before enabling file or command tools.

Open Agent mode

Request a small observable change

Start with a tiny application or one contained improvement. Require the agent to show changed files, tests, and any command that needs approval.

Try this prompt

Create a small accessible snake game in this workspace. Run its checks, inspect the result, and report the files changed and any remaining risk.

Inspect evidence before trust

A confident completion message is not proof. Review the diff, test output, browser behavior, and checkpoint or rollback route before accepting the work.

Open Work Board

Finish with an honest check

Complete the lesson only when you performed or deliberately inspected each observation—not simply because you read the page.