audit

Automated code audits for staged files run parallel subagents to check modified code against configurable project standards. You initiate a staged file audit to automatically validate code quality and enforce architectural rules before committing.

Parallel auditing works by deploying isolated subagents that evaluate multiple files simultaneously against defined criteria files. This allows the system to perform tiered analysis based on change volume, enabling efficient full or lightweight directory scans across large project structures.

You can enforce custom project coding standards by defining configurable criteria files in markdown for various code types. These files allow you to specify and automatically enforce architectural rules and quality benchmarks for components, pages, and services.

Tiered analysis is an audit mechanism that adjusts review depth based on the volume of code changes. It performs full audits for large modifications and lightweight audits for smaller scopes, ensuring efficient standards enforcement without unnecessary processing overhead.

You need to define criteria files before running a directory scan, as these markdown files provide the specific standards for components, pages, services, and overall structure. The automated audit uses these configurable criteria to evaluate your project files.