Code Review is Anthropic's official Claude Code/Cowork plugin that runs several specialized subagents in parallel over a pull request to surface bugs, security, and maintainability issues with confidence scores. It is used to automate PR code review while keeping false positives low.
Last 12 weeks · weekly buckets. New tools get more accurate as observation weeks add up.
Tools that are especially used together with this one. The higher the value, the better the fit.
Code Review MCP is Anthropic's official Claude plugin that runs several specialized subagents in parallel to analyze a pull request's changes from multiple angles. Independent reviewers each check CLAUDE.md compliance, likely bugs, git history context, prior review comments, and code comment accuracy, then the results are combined.
Each finding is scored on a 0-100 confidence scale, and by default only high-confidence issues at 80 or above are posted as comments, sharply cutting down false positives and review noise. It prioritizes high-impact problems across security, performance, maintainability, and correctness.
In BuildersgateGG's tool directory it is classified as a code-quality tool used in AI coding tools to automatically review pull requests or diffs and catch issues a human might miss.
It is Anthropic's official Claude plugin that runs multiple specialized subagents in parallel over a pull request to find bugs, security, and maintainability issues with confidence scores.
Each finding gets a 0-100 confidence score, and only high-confidence issues at or above the default threshold of 80 are posted, cutting down false positives.
It analyzes in parallel from several independent reviewer angles: CLAUDE.md compliance, likely bugs, commit history context, prior review comments, and code comment accuracy.
It acts as a first-pass automated review that filters high-confidence issues; final judgment and merging still rest with a human reviewer.