gvm-code-review

Automate code review by dispatching parallel subagents grounded in expert knowledge to scan for specific defect classes. This phased pipeline automatically checks code quality, security, and spec adherence to reduce manual review time and errors.

The best way to run comprehensive code quality checks is using a phased pipeline with parallel defect-class subagents. This approach scans for various defects simultaneously, ensuring the code meets required quality standards through expert grounding.

Yes, you need a structured codebase with clearly defined specifications to perform automated code quality checks. The review process also requires access to expert knowledge for grounding the subagents during the spec adherence scan.

Expert grounding works by using expert knowledge to guide parallel subagents during the code review process. This ensures the subagents scan for defects according to specific domain knowledge, providing comprehensive coverage for security and quality.

Yes, you can review code for security vulnerabilities and quality in parallel. The process dispatches parallel subagents to scan for specific types of defects, ensuring comprehensive coverage across various aspects of code quality and security.

You should avoid using automated subagents for code review if your project lacks a structured codebase with clearly defined specifications. Without expert knowledge to ground the subagents, the automated checks cannot effectively scan for defects.