Cursor Bugbot configurable effort levels for PR reviews

Summary

Cursor added configurable effort levels for Bugbot PR reviews (May 11): Default (optimized for efficiency), High (more reasoning time, more expensive), and Custom (natural language policy describing when to use each). Teams admins and Individual plan users can set the policy. Also shipped a new PR review experience in Cursor 3 — creation to merge in one place.

Implications

• Effort-level pattern originated in Claude Code (default/low/high/xhigh), spread to the API tier, now reaches automated PR review
• First PR review tool to expose cost-quality tradeoffs as a tunable parameter — teams can set natural language policies like “use high effort for security-sensitive files”
• Extends Cursor’s enterprise governance stack: model controls (May 4) → spend limits (May 4) → context usage (May 6) → PR review + parallel plans (May 7) → effort levels (May 11). Five features in eight days.
• Feeds: enterprise deployment battleground (Cursor governance dimension), agent lifecycle thread (effort as a pattern)