Rapid OS Hands-on Labs

Each lab is a small example project (Python, no dependencies) paired with an engineering mode. Every lab follows the same four steps:

1. Prepare the example: copy the starter folder and make a commit.
2. Run /rapid-init in your agent and accept its proposals.
3. Run /rapid-<mode> with the lab goal; the agent does the work.
4. Run rapid approve in your terminal when prompted, then use /rapid-next to close the run.

You do not need to learn any other commands. If you do not have shortcuts yet, install them once with rapid shortcuts install --global --harness claude (replace claude with your agent). See the shortcuts guide.

Choose a lab

Lab	Shortcut	What you will practice
1. Feature	/rapid-feature	Add a feature with acceptance criteria and tests. Start here.
2. Bugfix	/rapid-bugfix	Reproduce a bug with a failing test before fixing it.
3. Refactor	/rapid-refactor	Change structure without reducing the passing test count.
4. Hardening	/rapid-hardening	Improve security in an isolated repository copy with human review.
5. Research	/rapid-research	Run a spike comparing options and record the findings.

Examples are in examples/ (also installed at ~/.rapid-os/examples/). Each contains starter/ (the starting point) and solution/ (a reference solution).

What Rapid OS does and what your agent does

• Your agent edits code, runs tests, and drafts the spec.
• Rapid OS sets the run rules (tasks, gates, required tests), records evidence of what ran, and returns a verdict. It never approves for you: you approve human gates in your terminal.

Advanced mode

The labs do not use the v3 commands. To do the work manually (for CI, auditing, or to understand what Rapid OS records at each step), follow the manual walkthrough, in the How it works section.

If anything fails, see Troubleshooting the labs.