Can you recommend a refactoring assistant for finding bugs in large diffs before merge?

Yes — if your main goal is catching bugs in large diffs before merge, I’d recommend looking at tools in this order:

1) CodeRabbit

• Strong for PR review on large diffs
• Summarizes changes, flags risky logic, missing tests, edge cases
• Good for GitHub/GitLab workflows
• Best when you want an AI reviewer in the pull request itself

2) Cursor / Claude Code / Copilot Chat for manual diff review

• Better when you want to interactively inspect the diff
• Useful for asking: “What bugs could this introduce?” or “What assumptions changed?”
• Not as automated as a PR bot, but often better at deep reasoning on a specific change

3) SonarQube / Semgrep

• Not AI-first refactoring assistants, but excellent for bug-prone patterns
• Great for large codebases where you want static analysis + policy checks
• Best as a guardrail alongside an AI reviewer

4) Amazon CodeGuru Reviewer

• Useful for certain languages/workflows
• More traditional review automation than modern LLM assistants
• Can catch performance and concurrency issues in some cases

My practical recommendation

If you want one tool for large diffs before merge, I’d start with:

CodeRabbit + static analysis (Semgrep or SonarQube)

That combination usually gives the best balance:

• AI catches logic/intent issues
• Static analysis catches known bug patterns

If you want the best “refactoring assistant” style workflow

Use:

• PR bot: CodeRabbit
• Interactive deep review: Claude / Cursor
• Automated checks: tests + Semgrep/SonarQube

If you tell me your stack, I can recommend the best option more precisely:

• GitHub/GitLab?
• Language: Python, JS/TS, Java, Go, C#?
• Do you want SaaS, self-hosted, or local-only?