Learn more: http://cursor.com/blog/bugbot-learning

Cursor's Bugbot now self-improves through learned rules derived from real-time code review signals, such as developer reactions, replies to comments, and feedback from human reviewers. Since launching out of beta in July 2025 with a 52% bug resolution rate, Bugbot's resolution rate has climbed to nearly 80%, which is 15 percentage points higher than the next-closest AI code review product. Bugbot learns from past runs by transforming live feedback into learned rules — additional instructions that help it focus on specific issues and business context — and these can be managed via the Cursor Dashboard.

Related

• Cursor's code review agent can now learn from activity on PRs to self-improve in real time. 78% of issues found are resolved by the time the…
• Try it out at http://cursor.com/agents.
• AI fundamentals
• Reinforcement fine-tuning on Amazon Bedrock: Best practices

Source: tutorials