Cursor lets cloud agents build and host apps, gated by a pre-publish security review

Cursor says a user can start a cloud agent, tell it to build an app, and get back a live link. The agent scaffolds the web app, creates a Postgres database if one is needed, puts authentication behind Cursor SSO, and provisions hosting at a domain the user picks. It can deploy from new or existing repositories on Cursor Origin and GitHub, with GitLab support listed as coming. Cursor says the feature is available on all paid plans that have cloud agents enabled.

Before publishing, Cursor runs deterministic and agentic security checks aimed at problems such as secrets written into app files. The post says: "Nothing goes live until the checks pass and you explicitly publish the app." Team admins can see every app and secret on the team, add their own rules to the review, approve a publish the review blocked, and unpublish any app. Apps are shared behind Cursor sign-in, and individual plans can also share them publicly. Cursor says staff in sales, operations, legal, finance and engineering are already building internal apps this way. The post gives no usage figures and describes the checks only by example, and it is Cursor's own announcement, with no independent testing found.

Why it matters: Publishing is where an agent's output starts to be used by other people, and Cursor puts a review gate and an admin override at that point. If your team adopts this, the review rules and the list of who can approve a blocked publish become part of your harness. They are only as strong as the checks behind them, and Cursor has not said what those cover beyond its examples.