How many issues will the skill report?

The skill is configured to find a minimum of 3 and a maximum of 10 specific problems for every story or task it reviews.

What standards does the review check against?

The skill evaluates code quality, test coverage, architectural compliance, security vulnerabilities, and performance efficiency.

Does this skill provide automated fixes?

Yes, once the Adversarial Code Review skill identifies problems, it can automatically apply fixes to the code subject to user approval.

What is an adversarial code review?

An adversarial code review is a rigorous evaluation process where the AI agent is programmed to actively find flaws rather than just scanning for errors, ensuring that even subtle quality issues are addressed.

Can the skill be used for any programming language?

Yes, as a Claude Code skill, it leverages Claude's multi-language capabilities to review any codebase it has access to within the terminal environment.

Adversarial Code Review

Name: Adversarial Code Review
Author: PabloLION

byPabloLION

•

コラボレーションツール

Performs rigorous, adversarial senior developer code reviews to identify and resolve critical flaws in every story.

The Adversarial Code Review skill transforms Claude into a strict senior developer who refuses to accept 'looks good' as an answer. It systematically scrutinizes every story to find between 3 and 10 specific problems, challenging everything from architectural compliance and security to test coverage and performance. Built as part of the BMAD agile methodology, this skill ensures that no code reaches production without passing a high bar of excellence, while providing the convenience of automated fixes upon user approval.

主な機能

012 GitHub stars

02Comprehensive audits for security, performance, and architectural integrity

03Strict adversarial logic that prevents premature approval of PRs

04Automated remediation of identified issues with user confirmation

05Mandatory identification of 3-10 specific issues per code review

06Deep analysis of test coverage and code quality standards

ユースケース

01Enforcing architectural consistency across complex solo or team projects

02Ensuring production-grade quality in pull requests before merging

03Identifying hidden security vulnerabilities and performance bottlenecks

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add pablolion/bmad-plugin code-review

For use in Claude.ai and ChatGPT