Is this skill language-specific?

No, it is language-agnostic. It uses your project's context and framework files to adapt its review criteria to any programming language or stack.

When should I use this skill?

Use it when you finish a task, before creating a pull request, or anytime you want expert feedback on staged changes to ensure best practices are followed.

How does the persona-based review work?

The skill analyzes your diff and automatically selects relevant specialists, such as security or performance reviewers, alongside 'always-on' correctness and testing agents to provide a multi-dimensional perspective.

What is the P0-P3 severity scale?

It categorizes issues from P0 (Critical/Must Fix) to P3 (Minor/Discretionary), helping developers prioritize high-impact changes and security vulnerabilities before merging.

Can it fix the issues it finds automatically?

Yes, the skill includes a 'Standalone Fix Loop' that allows you to select specific severity levels for Claude to automatically address, commit, and then offer a re-review.

Multi-Agent Code Review

Name: Multi-Agent Code Review
Author: tmchow

bytmchow

•

Security & Testing

Conducts comprehensive, persona-driven code reviews using parallel AI sub-agents to identify bugs, security flaws, and architectural debt.

The iterative-engineering code-review skill transforms Claude into a sophisticated review team by spawning specialized sub-agents tailored to your specific code changes. By utilizing eight distinct reviewer personas—ranging from security and performance experts to maintainability specialists—it provides structured, deduplicated feedback categorized by severity (P0-P3). Whether used standalone before a PR or as part of an iterative development workflow, this skill ensures high code quality through automated intent discovery, rigorous confidence gating, and an integrated fix-and-re-review loop that streamlines the path to a clean merge.

Key Features

018 Specialized Reviewer Personas (Security, Performance, Logic, etc.)

02Integrated Fix Loop with interactive severity acceptance

03P0-P3 Severity Ranking with actionable merge-ready verdicts

04Automated Intent Discovery via branch and commit history analysis

0520 GitHub stars

06Parallel Sub-agent Execution for faster, multi-perspective feedback

Use Cases

01Automated cleanup of technical debt and maintainability issues.

02Pre-PR sanity checks to catch critical bugs and security vulnerabilities.

03Continuous feedback during iterative implementation of complex features.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tmchow/tmc-marketplace code-review

For use in Claude.ai and ChatGPT

Download Skill