adversarial-review

Name: adversarial-review
Availability: InStock
Author: Sendbird

Challenge design tradeoffs in local code changes with Claude Code.

161|28|Updated Apr 5, 2026

Search Tags:#claude-code #code-review #adversarial #review-automation #base-branch #working-tree #focus-text

Ready to run

npx skills add https://github.com/sendbird/cc-plugin-codex --skill adversarial-review-sendbird

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: adversarial-review
Source: https://github.com/sendbird/cc-plugin-codex/tree/main/skills/adversarial-review
Command: npx skills add https://github.com/sendbird/cc-plugin-codex --skill adversarial-review-sendbird

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Claude Code adversarial review helps you challenge design choices, tradeoffs, and risk assumptions in local code changes to surface issues early.

Core Features & Use Cases

Perform a design-challenging critique of diffs, branch changes, and risky migrations.
Preserve focus text and scope flags while evaluating architecture, safety, and maintainability tradeoffs.
Ideal for architecture reviews, high-risk refactors, and when a regular code review might miss deeper design implications.

Quick Start

Invoke adversarial review with optional flags to challenge the design of your local changes.

Frequently Asked Questions about adversarial-review

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

FAQPage Schema

What is adversarial code review and how does it challenge design tradeoffs?▼

Adversarial code review challenges design choices in local code changes to surface architecture issues early. It evaluates risk assumptions and maintainability tradeoffs across base-branch, working-tree, or auto scope scenarios to catch deeper design implications.

How do I run an adversarial review on local working-tree changes?▼

Invoke the adversarial review with explicit arguments to challenge local changes. Specify the target scope using --base, --scope, or --model flags to evaluate base-branch, working-tree, or auto scenarios, while preserving any user-provided focus text.

Can I focus an architecture review on specific code changes using scope flags?▼

Yes, you can target specific architecture questions by providing optional focus text alongside --base, --scope, and --model arguments. The review preserves your focus text and enforces explicit argument handling to evaluate the specified changes.

Does adversarial code review work for high-risk refactors and migrations?▼

Adversarial code review is ideal for high-risk refactors and risky migrations. It performs a design-challenging critique of branch changes to evaluate architecture, safety, and maintainability tradeoffs that regular reviews might miss.

When should I use an adversarial review instead of a regular code review?▼

Use adversarial review when a regular code review might miss deeper design implications. It is built to challenge design tradeoffs, risk assumptions, and architecture questions in local code changes before they become critical issues.