Bugbot adds PR review effort controls
Cursor now lets teams and individual Bugbot users choose how deeply the PR reviewer thinks, with default, high, and custom effort modes. High effort spends more time reasoning and, per Cursor, finds about 35% more bugs than default while keeping the same merge-time resolution rate.
This is a practical knob, not a flashy one, and that makes it more important: Cursor is turning code review quality into an explicit tradeoff between latency and bug yield.
- –High effort is the right default for high-risk code paths like infra and backend, where missed bugs are expensive
- –Custom effort is the more interesting piece because it makes review policy configurable in plain English, not just a binary toggle
- –The reported lift from 0.7 to 0.95 bugs per run suggests Cursor is pushing Bugbot toward a more selective but deeper reviewer
- –Usage-based billing gating the feature keeps it aligned with teams that already care about measurable review ROI
- –This strengthens Bugbot as a production code-review tool, not just an AI comment generator
DISCOVERED
1h ago
2026-05-11
PUBLISHED
2h ago
2026-05-11
RELEVANCE
AUTHOR
cursor_ai