parallel-review

name	parallel-review
description	Run parallel code review (quality, security, performance)

Parallel Code Review

Start multiple review agents in parallel that review code from different perspectives.

Workflow

Start 3 agents in parallel using the Agent tool:

Agent 1: Code Quality Review

Review the code for:
- Readability and naming
- DRY principle, unnecessary complexity
- Consistency with existing patterns in the project
- Missing error handling

Agent 2: Security Review

Review the code for:
- XSS, injection, unsafe data handling
- Auth/authz: Are permissions checked?
- Sensitive data in logs or client state?
- OWASP Top 10 relevance

Agent 3: Performance Review

Review the code for:
- Unnecessary re-renders (React)
- N+1 queries or missing pagination
- Bundle size impact (new dependencies?)
- Caching opportunities (TanStack Query staleTime, etc.)

Consolidate Results

Collect the three reviews and create a prioritized list:

1. Blocker - Must be fixed before merge
2. Important - Should be fixed, but can be a follow-up
3. Nice-to-have - Improvement suggestions

Note

This pattern is especially suited for greenfield projects or larger features. For small changes in existing projects, a single review is sufficient.