How does the decision log improve future reviews?

The decision log records accepted patterns and architectural choices, which the agents check before reviewing to avoid flagging established conventions as issues.

What happens if multiple agents find the same issue?

The workflow includes a consolidation phase that merges identical findings and credits all relevant reviewers, ensuring the final report is clean and concise.

Can I define my own specialized review agents?

Yes, the skill provides a framework for defining custom agent roles such as accessibility experts, documentation reviewers, or database optimization specialists.

How does this skill achieve true parallel execution?

It utilizes the Claude Code Task tool to launch multiple specialized sub-agents within a single message, allowing them to process the code diff concurrently rather than sequentially.

Is this skill compatible with monorepos?

Absolutely. It is particularly effective for large changesets in monorepos where different agents can focus on different sub-systems or layers of the stack simultaneously.

Parallel Code Review

Name: Parallel Code Review
Author: dgalarza

bydgalarza

•

Collaboration Tools

Orchestrates multiple specialized AI agents to perform concurrent security, architecture, and performance reviews on code changesets.

The Parallel Code Review skill optimizes the code review process by launching specialized sub-agents simultaneously, allowing for deep, multi-perspective analysis in the time it takes for a single review. It addresses the common bottleneck of sequential analysis by running security, architecture, and performance experts in parallel, then consolidating their findings into a single, prioritized report. With built-in decision tracking to prevent redundant feedback and language-specific adaptations for frameworks like Rails and Django, this skill is essential for teams maintaining high code quality and security standards in fast-moving development environments.

Key Features

01Concurrent execution of multiple specialized review agents

02Automated deduplication and consolidation of review findings

03Persistent decision logging to prevent repetitive suggestions

04Severity-based reporting for actionable feedback

05Language-specific templates for Rails, Django, and Node.js

0613 GitHub stars

Use Cases

01Performing rigorous security and architecture audits on large feature branches

02Standardizing organizational best practices through persistent decision tracking

03Reducing review turnaround time by parallelizing multi-domain analysis

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dgalarza/claude-code-workflows parallel-code-review

For use in Claude.ai and ChatGPT

Download Skill