Cursor Bugbot Autofix Closes Code Review Loop with Automated Repairs

Cursor

Feb 26, 2026 · Updated Apr 25, 2026

Cursor's Bugbot can now automatically fix the issues it finds in PRs, not just flag them. Autofix spawns cloud agents in isolated VMs to test and resolve bugs, with 35%+ of proposed changes getting merged directly into the base PR.

Cursor extended its code review agent with Bugbot Autofix, moving from flagging issues to automatically fixing them. When Bugbot identifies a bug in a PR, Autofix spawns a cloud agent in an isolated virtual machine to test and repair the issue, then proposes a fix you can merge with one click or accept via @cursor push in a PR comment.

The numbers back the momentum: resolution rate jumped from 52% to 76%, average issues per run nearly doubled, and 35%+ of Autofix changes are merged rather than just reviewed. Bugbot is also catching more bugs while flagging fewer false positives.

Autofix is out of beta and available to all Bugbot users. Enable it in your Bugbot dashboard - once on, every PR Bugbot reviews will include proposed fixes alongside the issue report.

View the full update on cursor.com

Cursor

@cursor_aiFeb 26

Cursor can now automatically fix issues it finds in PRs with Bugbot Autofix. https://t.co/4kZAy7DFel

View on X

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

See all AI news & updates from Cursor →

Keep reading

Cursor Bugbot Now Learns From Pull Request Feedback to Self Improve

Cursor's code review agent, Bugbot, can now automatically generate repository-specific rules by analyzing developer reactions and human reviewer comments on pull requests. This real-time learning loop has increased the agent's issue resolution rate to 78%, significantly reducing false positives.

ClaudeMar 28

Anthropic Claude Code now autonomously fixes CI failures and PR comments

Anthropic introduced a cloud-based auto-fix feature for Claude Code that monitors GitHub Pull Requests to resolve CI errors and reviewer feedback. This shifts the agent from a local tool to an autonomous background worker, allowing developers to submit code and walk away while the AI handles the iterative review cycle.

Cursor Publishes CursorBench, Its Internal Agentic Coding Evaluation Methodology

OpenAIMar 15

Cursor Publishes CursorBench, Its Internal Agentic Coding Evaluation Methodology

Cursor published CursorBench, its internal eval suite that scores models on real coding agent tasks from actual developer sessions. Public benchmarks struggle to differentiate frontier models reliably — CursorBench produces more separation where it matters most.