Cursor launches Rollouts and Security Review, two bots that watch deploys and read every pull request

A coding agent can now open more pull requests in a day than a small team can read. Cursor released two bots today that read every one of them: Rollouts watches every change as it deploys and reports its health per environment, and Security Review posts one comment on every pull request naming bugs an attacker could use. Both are available today on Cursor Teams and Enterprise plans.

What Rollouts does

When a pull request opens, Rollouts reads the diff and the systems it touches, then writes a monitoring plan as a PR comment. The plan lists the risks it identified, the effect the change is meant to have, the signals it will check in your logs and metrics, and any gaps in instrumentation that would make the change hard to verify. Cursor says the plan is editable in the PR, and Rollouts then uses your version.

Rollouts wakes on deploy events for the change's commit and runs the plan against your logs, metrics, and traces. Each environment is tracked separately, so a change can be verified in staging and flagged in production. When Rollouts detects a regression, Cursor says it names the change it suspects, notifies the author, and can either open a revert pull request for review or hand the finding to a cloud agent for a fix. Rollouts does not merge or roll back on its own today. Cursor calls Rollouts a rebuilt version of Firetiger Change Monitors, running on its Bot Development Kit. Integrations at launch cover Origin or GitHub for source control, your continuous delivery system for deploy events, and Datadog and other telemetry providers for signals. Feature flag integration is described as coming soon.

What Security Review looks for

Security Review reads every pull request in the context of the codebase and posts one review comment reporting exploitable bugs. Cursor keeps style and quality on its existing Bugbot, and draft pull requests are skipped. Cursor says Security Review looks for injection across SQL, command, and template surfaces, along with authentication and authorization bypasses, including checks that a refactor stopped running. It also flags secrets and credentials committed to source, server-side request forgery and unvalidated redirects, unsafe deserialization, and dependency changes that introduce known vulnerabilities.

Each finding carries a severity, the attack path, and a proposed fix. Cursor says a finding dismissed with a reason is not raised again on that pull request. Teams can add codebase rules, such as which client all external calls must go through or which tables must never be queried from a request handler, and Cursor says Security Review enforces those rules on every pull request.

For a team already running Cursor's coding agents, the practical change is that pull-request review and post-deploy checks are now two more bots inside the same tool. A pull-request comment naming an exploitable bug, with the attack path described and a proposed fix attached, is easier to act on than a scanner report in another dashboard. Whether the false-positive rate is low enough to leave the bot on for every pull request is what teams will learn from the ten-day trial credits Cursor is including: Cursor says the credits cover "roughly 50 changes" on Teams and "roughly 500" on Enterprise. Cursor has not published its own false-positive numbers.

Source

• Rollouts and Security Review, Cursor changelog, 23 September 2026