The project has not shipped a public release yet. Security fixes land on main and are picked up by anyone running a checkout. Once releases start, the latest minor line is supported and older lines get fixes at maintainer discretion.

Email the maintainers or open a GitHub security advisory on this repository. Do not open a public issue for something exploitable.

Include what you found, how you confirmed it, and what a fix would need to change. You will get an acknowledgment within 3 days and a triage decision within 10 days.

In scope:

• the control plane (API, web, worker, gateway, runner agent);
• the content SDK and any code path that renders contributor content;
• the packaged Compose/Helm deployments;
• anything that lets a lab container reach the host, control plane, or another tenant's lab.

Out of scope:

• the deliberately vulnerable room targets themselves. Exploiting them is the entire point of the platform. Findings that only affect a room container from inside its own lab network are not vulnerabilities.
• reports from automated scanners without a working proof.

Test only against your own installation. Never attack infrastructure you do not own, and never expose intentionally vulnerable targets to networks you do not control.

We coordinate disclosure: fix first, advisory second, credit to the reporter unless they prefer anonymity.