Does it check for specific security vulnerabilities?

Yes, it is specifically configured to look for business logic flaws, race conditions (TOCTOU), information disclosure, and authorization bypasses like IDOR or path traversal.

What is the difference between Review Opus and standard reviews?

Review Opus utilizes the high-reasoning Claude Opus model to focus on deep architectural flaws, subtle edge cases, and long-term maintainability that faster models might miss during a quick scan.

How does the skill determine if a review passes or fails?

It follows strict decision rules where any error-level severity, security issue, or poor test quality automatically triggers a 'needs_changes' status in the final report.

Can this skill review my project plan before I start coding?

Yes, if a refined plan exists without implementation results, Review Opus automatically switches to Plan Review mode to evaluate technical approach soundness and potential failure modes.

Where are the review results stored?

All findings, including architectural notes and categorized issues, are written to a standardized .task/review-opus.json file for easy tracking.

Review Opus (Deep Review)

Name: Review Opus (Deep Review)
Author: Z-M-Huang

byZ-M-Huang

•

Security & Testing

Performs rigorous architectural and security reviews of code and project plans using advanced Claude Opus reasoning to identify subtle bugs and technical debt.

Review Opus is a specialized Claude Code skill designed for high-stakes code and plan validation, leveraging the superior reasoning of the Opus model. It acts as a secondary, thorough layer of analysis following initial reviews, focusing on complex architectural integrity, long-term maintainability, and hidden security vulnerabilities like race conditions or authorization bypasses. By analyzing both refined plans and implementation results, it provides a comprehensive assessment of technical soundness and test quality to ensure production-grade software delivery through standardized JSON reporting.

Key Features

01Dual-mode review logic for both implementation plans and final code results

02Automated decision engine that enforces strict quality standards for approval

0311 GitHub stars

04Deep architectural analysis focusing on long-term maintainability and technical debt

05Rigorous security auditing for business logic flaws, race conditions, and IDOR vulnerabilities

06Comprehensive test quality assessment using FIRST principles and coverage analysis

Use Cases

01Enforcing high test standards and maintenance practices in enterprise-grade repositories

02Validating complex technical plans before implementation to prevent costly architectural mistakes

03Conducting high-level security audits on critical code paths to catch subtle vulnerabilities

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add z-m-huang/claude-codex review-opus

For use in Claude.ai and ChatGPT

Download Skill