Cursor Bugbot Review Guide

The figures in the table come from Cursor's June 2026 changelog. Check them before you quote them, because BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition.'s speed and cost can change with later updates and its billing can change by plan.

Cursor's release notes and announcements say what changed. They don't say how a team should act on it. The steps below are our suggested order, and the next section lists what to watch out for along the way.

1. 1Run BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. on real PR diffs, alongside your tests and code owners.
2. 2Use /review locally before you push a risky or security-sensitive change.
3. 3Set model blocklists, effort and incremental review so findings match your team's risk.
4. 4Treat findings as review evidence. A person still decides what merges.

Each item below is either a fact about the release, such as its beta status or plan access, or our advice for using it safely. Plan around each one, and recheck the facts when Cursor ships an update.

• A clean BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. run is not merge approval. It means Bugbot flagged nothing, and your tests and code owners still have to pass the change.
• Tune noisy findings with repo rules and effort levels instead of ignoring the tool.
• Recheck pricing and effort settings, because BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. billing and usage can change by plan.
• BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. is available on Premium and higher, Teams and Enterprise, and on the legacy Pro, Pro+ and Ultra plans. Free, Plus and the legacy Start plan don't include it.

BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. is Cursor's AI code reviewer. It runs on pull requests in GitHub, GitLab, Bitbucket, Azure DevOps and Origin. In one Cursor talk, the presenter called its signal-to-noise ratio high: consistently, 70 to 80% or more of the issues it raises get resolved. It flags problems such as exposed keys and proposes fixes.

Settings to tune

• Which repos BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. reviews.
• Effort level: how long BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. reasons during a review, on usage-based Bugbot plans only.
• Automatic review on every PR, or a manual trigger only.
• Autofix on, or off so that BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. only reports issues.

Risk-rated PRs

Pair BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. with an automation that classifies each new PR by risk. Low-risk PRs, such as a title change, get approved automatically and merge with little friction. High-risk PRs need human review, and the automation assigns reviewers by using git blame to find the people who touched that code.

For leadership, the presenter showed a 30-day dashboard of PRs reviewed, issues resolved and automatic fixes merged. You can drill into each repo and break the results down by severity: high, medium or low. It is the main view for judging the ROIReturn on Investment. The value gained versus what it cost, the language an economic buyer funds deals in. Press Enter for the full definition. of AI review at scale.

Make BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. the merge gate, and set how strict the gate is by risk. One finance team builds accounting apps in a dedicated section of the company monorepo. It runs Bugbot as an AI reviewer on every PR, then lets each PR's risk decide who else has to look.

It proves that a machine pass found nothing it recognizes. It doesn't prove the change is the right one, so it isn't merge approval.

The clip below shows the opposite case, a run that comes back with findings.

What a clean run can't check is intent. Any reviewer, human or bot, can judge whether a change looks safe. Only the person who asked for the change knows whether it is the change they wanted. That gap doesn't shrink as volume grows, so treat a clean run as the starting point of a review.

Accountability doesn't move either. When a merged change breaks production, the person who clicked merge owns it.

The win is narrow and still worth having. You start reading a diff that has already been swept for mechanical problems. That saves the first stretch of a review and none of the judgment at the end of it.

Tune it in the first week. A reviewer who has learned to skim past the bot's comments is worse off than one who never had it, because those comments now fill the slot where reading used to happen.

You don't need to find a perfect setting. Teach BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. two or three specific things about this repo, then leave the rest alone.

Two levers do most of the work. The effort level, on usage-based BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. plans, sets how long Bugbot reasons over a review. Repo rules teach it conventions that a general-purpose reviewer has no way to know. For a finding that keeps coming back wrong, a rule is the better lever. Add one after you have seen the same finding twice, and keep it short so it doesn't dilute the other rules.

Leave incremental review on. It is the default, and it reviews only what changed since the last BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. pass, so a second push doesn't re-review code the first pass covered. Findings that repeat on every push make a bot feel noisier than any single wrong finding does.

If findings still read as noise after both levers, look at the size of the pull request before you blame the reviewer. A very large change gives a bot more surface to be pedantic about. Stacked diffs shrink that surface without touching a single setting.

Each one runs at a different moment. /review runs before you push, BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition. runs on the pull request, and a Vulnerability Scanner runs on a schedule against the code already in the repo.

The local pass is the cheapest habit of the three. /review and /review-security run in Cursor 3An older release label for Cursor's multi-agent interface; current product guidance calls the active surface the Agents Window. Press Enter for the full definition..7 and later, in the web app at cursor.com/agents and in the Cursor CLICursor's command line: the full agent, all modes and models, in the terminal and pipeable into scripts and CI. Press Enter for the full definition.. /review-security compares your branch against the default base branch and includes uncommitted work, so name the branch when your base is not the default.

The scheduled scanner is a separate agent type from BugbotCursor's automated PR reviewer that posts inline findings and can push fix commits from isolated VMs. Press Enter for the full definition.. Cursor's Security Agents come in two kinds. The Security Reviewer takes git triggers and checks pull requests before they merge. The Vulnerability Scanner takes cron triggers and looks for problems that pull request review never saw. Both bill to the team's usage pool under a shared team service account, so they don't count against any one person's usage.

Their setup differs. A Security Reviewer needs at least one tool or MCPModel Context Protocol. A standard that lets an AI agent pull in context from outside the repo, like Jira tickets or internal docs. Press Enter for the full definition. before you can save it. A Vulnerability Scanner saves without one and reports its findings to a list of flagged vulnerabilities. Both run on Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition., except a team's Security Reviewer triggered by a pull request event, which runs on Cursor-managed hosting.

On a small team, run the local pass and stop there. A scheduled scanner pays off once the codebase is older than the review habit, meaning there is code that nobody now reading pull requests ever reviewed.