pr-review

Perform a diff-based code review by analyzing changed files to identify high-confidence issues, routing checks across code quality, test coverage, error handling, type design, and comment quality. Produce severity-ranked findings with file:line locations and recommended actions.

Prioritize critical fixes in PR reviews by reporting Critical, Important, and Suggestion issues with confidence scores. Deduplicate overlapping findings and favor precision over noisy recommendations to reduce regressions and review noise.

Diff routing selectively loads methodologies for test coverage and error handling based on the specific changed files in the pull request. This targeted approach ensures only relevant code quality dimensions are evaluated.

Use CLAUDE.md rules to calibrate code review findings by respecting project-specific configurations. Deduplication and calibration rules merge overlapping findings and favor precision over noisy recommendations for your pre-merge report.

Identify silent failures and type design regressions by running diff-based code reviews on a feature branch. Generate an actionable pre-merge report highlighting security regressions, test gaps, and type issues with mandatory file:line locations.

Automated diff analysis focuses only on changed files to produce high-precision code reviews. It targets high-confidence issues and deduplicates findings, but does not evaluate the entire repository context outside the current diff.