prp-feature-review-agents

Deep code review for large software packages is performed by spawning parallel, specialized agent subprocesses. Each agent operates with a fresh context window to conduct thorough, code-level analysis without hitting memory limits.

Code quality and security are analyzed simultaneously during a feature review by dispatching specialized agents in parallel. Distinct agents handle specific focus areas including performance, accessibility, dependencies, and observability.

The best way to review a complex codebase for iterative improvement is using multi-agent analysis. Specialized agents delve into all files within the package, enabling continuous refinement through iterative review-fix cycles.

Multi-agent feature review is explicitly designed for large and complex codebases. It handles scale by giving each specialized agent a fresh context window, allowing deep exploration of all package files.

To start a feature review on a specific package folder, trigger the analysis on the target directory. The system automatically spawns parallel agent subprocesses to evaluate code quality, security, and performance.