Can it handle conflicting feedback from different reviewers?

Yes, it implements a Critic phase that cross-validates findings across reviewers to identify contradictions, resolve severity levels, and present a unified recommendation to the user.

Does this skill provide specific code locations for its findings?

Every finding is required to follow a strict 'No Vague Findings' policy, providing specific file paths, line numbers, and concrete evidence found through tools like Grep and Read.

How does the Zen Reviewer skill choose which experts to use?

It uses a selection algorithm that matches changed files against specific patterns (e.g., .sql for Database experts, .jsx for Frontend experts) and analyzes diff content for relevant keywords like 'auth' or 'crypto'.

How are security issues prioritized in the review process?

The Security Expert is automatically included in the review process if any security-sensitive files or keywords are detected, and security findings are often given high priority in the final report.

Zen Reviewer Coordinator

Name: Zen Reviewer Coordinator
Author: B16B1RD

byB16B1RD

0•

コラボレーションツール

Orchestrates specialized expert agents to perform comprehensive, high-quality pull request reviews based on domain-specific expertise.

The Zen Reviewer skill acts as a central coordinator for automated pull request reviews within the Claude Code environment. It leverages a specialized selection algorithm to assign expert agents—such as Security, DevOps, Frontend, and Database specialists—based on changed file patterns and content analysis. By implementing a Generator-Critic pattern, it ensures high-quality findings, cross-validates conflicting suggestions, and produces a unified, actionable report. This skill is essential for teams looking to automate rigorous code reviews while maintaining high standards across their entire technical stack.

主な機能

010 GitHub stars

02Evidence-Based Findings: Enforces a strict policy against vague feedback, requiring agents to provide concrete evidence via investigation tools.

03Automated Expert Selection: Dynamically assigns specialized reviewers based on file patterns and content analysis.

04Generator-Critic Integration: Enhances review quality by having a critic agent validate and reconcile findings from multiple experts.

05Unified Reporting: Aggregates domain-specific analysis into a single, structured summary with clear severity levels.

06Cross-Validation Logic: Detects contradictions between reviewers and increases severity for issues flagged by multiple agents.

ユースケース

01Automating multi-disciplinary PR reviews in complex, full-stack repositories.

02Enforcing domain-specific best practices for security, API design, and database migrations.

03Reducing manual review overhead by pre-validating code changes against specialized expert checklists.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add b16b1rd/cc-zen-workflow reviewers

For use in Claude.ai and ChatGPT

Download Skill