code-review-pipeline

Multi-agent code review automation coordinates an orchestrator and multiple auditor agents to evaluate code changes. This architecture runs sequential audits for requirements and correctness, followed by parallel evaluations of quality dimensions to ensure thorough evaluation.

The multi-agent code review pipeline structures its workflow using a fixed architecture of an orchestrator, seven auditors, and a final synthesizer. It executes sequential audits for requirements and correctness first, then runs parallel evaluations across quality dimensions.

Standardize local code reviews by running the /ReviewLocal command to load a fixed multi-agent audit workflow. This process applies sequential and parallel auditing phases to evaluate requirements, correctness, test coverage, maintainability, and performance consistently.

Yes, you can use automated code review auditing for local changes before committing by running the /ReviewLocal slash command. This initiates a session applying the standard multi-agent audit workflow to evaluate your uncommitted codebase modifications.

An automated multi-agent review pipeline evaluates code quality dimensions including test coverage, maintainability, testability, performance, and extensibility. These dimensions are assessed in parallel during the auditing process following the initial sequential requirements and correctness checks.

Per-auditor Lessons Learned patterns improve automated code review workflows by capturing continuous improvement insights within each auditor agent. This enables the multi-agent architecture to refine its sequential and parallel auditing evaluations over successive code reviews.