review-opus

Analyze architecture, subtle bugs, and test quality from Sonnet review outputs.

Updated Jan 16, 2026
One-click install
npx skills add https://github.com/cskwork/llm-review --skill review-opus
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: review-opus
Source: https://github.com/cskwork/llm-review/tree/main/plugins/claude-codex/skills/review-opus
Command: npx skills add https://github.com/cskwork/llm-review --skill review-opus

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Thorough, end-to-end reviews identify architecture flaws, subtle bugs, and weaknesses in test coverage that quick checks miss, helping teams ship more reliable software.

Core Features & Use Cases

  • Plans: evaluate planned changes by inspecting .task/plan-refined.json for soundness, risks, and maintainability.
  • Code: analyze changed files via .task/impl-result.json and assess architectural integrity, edge cases, and test quality.
  • Output: generate a structured .task/review-opus.json with findings, issues, and architectural notes to guide remediation.

Quick Start

Invoke the opus skill after a Sonnet review to perform a deep analysis of the latest code changes.

Frequently Asked Questions about review-opus

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I perform a deep architecture review on my code changes?

To perform a deep architecture review, this Skill analyzes code changes via the .task/impl-result.json file, assessing architectural integrity, edge cases, security implications, and test quality, then outputs structured findings to .task/review-opus.json.

What is a deep code review and how does it differ from quick checks?

A deep code review identifies architecture flaws, subtle bugs, and weaknesses in test coverage that quick checks miss. It evaluates edge cases, security implications, and maintainability to help teams ship reliable software.

Can I use this to review planned changes before implementation?

Yes, you can review planned changes. When the .task/plan-refined.json file is present, the Skill evaluates the plan for soundness, risks, and maintainability before code is written.

Do I need a Sonnet review before running this deep analysis?

Yes, you invoke this opus skill after a Sonnet review. It is designed to perform a deep analysis of the latest code changes following that initial review process.

How do I get structured output from a security and maintainability review?

The Skill generates a structured .task/review-opus.json file containing a summary of findings, issues, and architectural notes to guide your remediation efforts.

What are the limitations of using automated deep reviews for testing quality?

Automated deep reviews analyze changed files to assess test quality and architectural integrity, but they rely on the presence of .task/impl-result.json or .task/plan-refined.json files to function properly.