compound-review

To run a multi-perspective code review, invoke the workflow on your pull request diff to coordinate four always-on reviewers and conditional specialists, generating a structured report of prioritized findings.

A multi-persona code review coordinates four always-on reviewers focused on correctness, maintainability, testing, and project-standards, plus conditional specialists for security or performance, catching issues single-perspective reviews miss.

Yes, automated code review applies to diffs from ongoing work across frontend and backend changes, activating conditional specialists based on the specific content of your diff to evaluate code quality before merging.

The best way to get prioritized code review findings is to run a multi-perspective review after implementation and before committing, which outputs a final report categorizing issues into P0 through P3 severity levels.

You need conditional specialists in a code review workflow when your diff content touches specific areas like security, performance, or API contracts, triggering their activation alongside the four always-on reviewers for comprehensive evaluation.

Automated code review supplements rather than replaces manual quality assurance by providing structured evaluation across correctness and maintainability personas, delivering a prioritized P0-P3 report to guide human remediation efforts before merging.