code-review-basic

Automated code review validates unstaged git diff changes against project guidelines defined in CLAUDE.md. It checks multi-language repositories for conventions, error handling, and security, outputting structured reports with severity scores and concrete fixes.

Issue scoring assigns a 0-100 severity value to detected code quality issues during review. This scoring system helps prioritize bugs, security vulnerabilities, and style deviations by generating actionable lists with file paths, line numbers, and CLAUDE.md rule references.

Reviewing a pull request for security and testing involves validating code changes against project conventions. The process identifies missing error handling, logic errors, and security issues, then generates a structured report with line numbers and specific fixes.

Automated code review supports multi-language repositories by validating changes against project guidelines. It applies to diverse review scenarios including imports, naming conventions, error handling, and testing practices without relying on language-specific component dependencies.

To reduce false positives in code review, validate unstaged changes strictly against established project guidelines in CLAUDE.md and report only high-confidence issues. This ensures consistency and compliance while minimizing manual review effort.