Cursor shipped a small but useful update to Bugbot, its AI-powered pull request reviewer: Teams admins and Individual plan users can now set the effort level Bugbot applies when reviewing PRs. There are three configurations to choose from, giving users direct control over how deep the AI digs into their code changes.
The feature addresses a real friction point with AI code review tools. A one-size-fits-all review intensity wastes time — thorough analysis on a trivial dependency bump is noise, while a shallow pass on a complex refactor misses bugs. Three distinct effort levels let developers match review depth to the actual risk of a given PR.
TIP
If you’re on a Teams plan, the admin sets the default effort level for the whole org. Individual plan users control it themselves. Check your Bugbot settings to make sure the default matches how you actually want reviews handled.
The changelog entry is sparse on specifics — Cursor doesn’t name the three levels or detail exactly what changes between them (number of passes, types of issues flagged, context window usage). That matters because “effort level” could mean anything from comment verbosity to actual analysis depth. Worth testing each setting on a real PR to see what you’re getting.
If you use Bugbot on your repos, go toggle the setting now. Start with the highest effort level on a PR you’ve already merged and compare its output against the issues you actually caught manually. That tells you whether Bugbot at max effort finds things you miss — or just generates more noise.