What methodology does this skill use?

It is based on Gregory Brown's 'Effective Code Reviews' (2015), emphasizing an outside-in approach that validates the problem and build before diving into the code.

What are the risk levels assigned by the skill?

Changes are categorized as Low, Medium, or High risk based on whether they touch shared code, modify existing tests, or involve sensitive areas like security and auth.

Does it actually run the code it reviews?

It verifies build status and logical runability using your project's tools, but it explicitly defers runtime user-experience verification to a human reviewer via a checklist.

How does this differ from standard linter or static analysis checks?

While linters focus on syntax and style, this skill evaluates logical intent, risk profiles, and testing adequacy, starting with whether the problem is actually solved.

Can I customize the review rules for my specific project?

Yes, the skill includes a setup phase that reads your existing project rules and allows you to configure specific CI commands, naming conventions, and output formats.

Code Review Expert

Name: Code Review Expert
Author: Agilefreaks

byAgilefreaks

0•

コラボレーションツール

Conducts systematic, risk-driven code reviews based on industry best practices to ensure architectural integrity and project alignment.

This skill implements an outside-in, risk-driven code review methodology inspired by Gregory Brown's 'Effective Code Reviews'. It shifts the focus from superficial style checks to deep structural analysis, beginning with problem validation before any code is read. The skill guides Claude through six rigorous phases—ranging from build verification and test auditing to the categorization of changes (such as bug fixes versus refinements)—to provide high-signal feedback, assign risk levels, and generate actionable checklists that highlight exactly what a human reviewer needs to verify manually.

主な機能

01Problem-first validation to ensure changes solve the intended business goal

02Context-aware coverage assessment for bug fixes, new features, and refinements

03Standardized risk assessment reports including detailed human reviewer checklists

040 GitHub stars

05Risk-based test auditing that flags suspicious modifications to existing test suites

06Automated build and runability verification to prevent reviewing broken code

ユースケース

01Standardizing code quality feedback across a distributed engineering team

02Auditing legacy code refinements where the potential blast radius is high

03Reviewing complex Pull Requests to identify architectural risks and logical flaws

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add agilefreaks/claude-skills code-review

For use in Claude.ai and ChatGPT

主な機能

01Problem-first validation to ensure changes solve the intended business goal

02Context-aware coverage assessment for bug fixes, new features, and refinements

03Standardized risk assessment reports including detailed human reviewer checklists

040 GitHub stars

05Risk-based test auditing that flags suspicious modifications to existing test suites

06Automated build and runability verification to prevent reviewing broken code

ユースケース

01Standardizing code quality feedback across a distributed engineering team

02Auditing legacy code refinements where the potential blast radius is high

03Reviewing complex Pull Requests to identify architectural risks and logical flaws