code-reviewer

A read-only code review targets a specified file or directory to identify security, logic, and concurrency issues without modifying project code. You simply run the tool against the target path, and all generated reports are safely written to a dedicated output directory.

Yes, you can audit changed files by scoping the code review to a git diff. This pre-merge audit identifies security vulnerabilities, dependency CVEs, and CI/CD risks specifically within your recent changes.

A full code audit scans across 10 dimensions including security, concurrency, performance, error handling, dependencies, architecture, and testing. It executes role-based scanning to produce deduplicated findings with required file-and-line evidence.

No, the code review tool enforces strict read-only behavior during analysis and never modifies your project code. It safely writes all structured review reports and deduplicated findings exclusively to a dedicated directory.

You can check for dependency CVEs and CI/CD risks by running a targeted code health check. The review process identifies dependency vulnerabilities and documentation or CI risks, providing a structured report with severity metadata.

A quick scan offers a fast conversation-based review, while a full audit performs a thorough multi-role scan with session resume capabilities. Both modes generate structured, actionable reporting with deduplicated findings and file evidence.