Guide · Agent skill
How to use the Code Reviewer agent skill: install and run it
Updated · written with AI from the project's own documentation and checked against our catalogue data
The code-reviewer skill is a set of instructions and report templates that teaches an AI agent to review code the way a senior engineer would. It is for developers who want a second pair of eyes on a pull request, and for teams running quick quality or refactoring audits. After this guide you will know what the skill does, how to install it, how to ask for a review in plain language, and what to check in the output.
Code ReviewerJeffallan/claude-skills · MIT Download the skill Quick start
- 1Download the ZIP. Download the code-reviewer skill ZIP from this page.
- 2Unzip into a skills folder. Unzip it into ~/.claude/skills/ for personal use, or into .claude/skills/ in your project for team use.
- 3Or upload in the Claude apps. In the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.
- 4Describe the review you want. Ask in plain language, for example to review a pull request or run a code quality check, and include the goal of the change.
- 5Check the summary and intent. Confirm the agent's one-sentence summary of the change matches what the author intended before trusting the rest.
- 6Work through the prioritized report. Address critical issues first, then major and minor ones, and answer the questions for the author.
What the code-reviewer skill makes the agent do
An agent skill is a folder of instructions that an AI agent loads when a task matches its description. The code-reviewer skill is described as analyzing code diffs and files to find bugs, security vulnerabilities such as SQL injection, XSS and insecure deserialization, code smells, N+1 queries, naming issues, and architectural concerns.
The skill works in a fixed order. The agent first summarizes what the change is trying to do in one sentence, then reviews architecture, code quality, security, performance and test coverage. It finishes with a report that groups issues into critical, major and minor, and includes praise for good patterns and questions for the author.
- Summary of intent and overall assessment
- Critical issues that must be fixed before merge
- Major issues that should be fixed
- Minor issues that are nice to have
- Positive feedback on specific patterns
- Questions for the author
- Verdict: Approve, Request Changes or Comment
Install the code-reviewer skill
Download the ZIP for the code-reviewer skill from this page and unzip it. You can then place the folder in one of two locations: your personal skills folder at ~/.claude/skills/, or a project-specific folder at .claude/skills/ inside your repository.
If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills.
- Personal install: unzip into ~/.claude/skills/ so it is available in every project.
- Project install: unzip into .claude/skills/ inside the repository so teammates who clone it get the same skill.
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP.
Trigger the skill with natural requests
You do not need a special command. The skill's description lists phrases that should activate it, such as PR reviews, code quality checks, refactoring suggestions, review code and code quality. Describe what you want in ordinary language and mention the code or change you care about.
Giving the agent the reason for the change helps. The skill asks the agent to summarize intent first, so a short description of the problem being solved produces better feedback.
- Review this pull request before I merge it. The goal is to add rate limiting to the login endpoint.
- Do a code quality audit of the payments module in this repo.
- Look for refactoring opportunities in src/orders/.
- Check this file for SQL injection or XSS risks.
- Review the code in this diff for N+1 queries and missing tests.
Read the review report
The skill uses a report template so that reviews look consistent across pull requests. Start with the summary, then work through critical issues first. Each issue should include a specific suggestion, and the report is meant to be read in priority order rather than file order.
Treat the verdict as advice. The skill's own wording is that a human should make the merge decision.
- Critical: bugs, security problems and data loss risks. Fix before merge.
- Major: performance, design and maintainability problems. Should be fixed.
- Minor: naming and readability. Nice to have.
- Questions for author: points the agent could not settle from the code alone.
Know what the skill needs and its safety limits
The skill lists no required dependencies. It is documented as not running scripts and not using the network. Its allowed tools are Read, Grep and Glob, which means it reads files and searches code. The skill mainly provides review guidance and report templates; it reads code rather than changing it.
Because it is a broad single-pass review, it can miss issues. The skill itself points to the security-reviewer skill for deeper security work, and the skill description notes it complements test-master for test-focused work. It is not a replacement for a dedicated security audit or penetration test.
- Requirements: none listed.
- Runs scripts: no.
- Uses network: no.
- Allowed tools: Read, Grep, Glob.
- Does not modify your code.
- Give it context: unexplained changes produce weaker results.
Tips for better reviews
Ask for one kind of review at a time when the change is large, for example security first and then performance. Point the agent at the specific files or diff. Where a linter or formatter is configured, the skill is designed not to nitpick style, so keep style checks in your tooling.
When the author has explained a non-obvious choice in comments, the skill is designed to acknowledge that reasoning before suggesting an alternative.
- Include the pull request description or a one-line goal in your request.
- Name the files or directories to focus on.
- Ask for the report template output so you get the standard sections.
- Use the questions for author section to start discussion with your team.
What people use it for
- Second-pair-of-eyes review of a pull request before merging.
- Quick code quality audit of a module before a release.
- Spotting N+1 query patterns in ORM code.
- Checking code for hardcoded secrets, SQL injection or XSS risks as a first pass.
- Finding refactoring candidates such as magic numbers and long functions.
Frequently asked questions
How do I install the code-reviewer agent skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use or into .claude/skills/ in a project. In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.
Does the code-reviewer skill require anything else to run?
The skill lists no required dependencies. It does not run scripts and does not use the network, and its allowed tools are Read, Grep and Glob.
Is the code-reviewer skill free?
The skill is released under the MIT licence. This page does not list pricing, so check the licence and the source repository for the terms that apply to you.
Will the skill change my code?
No. The skill mainly provides review guidance and report templates, and it reads code rather than changing it.
Can the code-reviewer skill replace a security audit?
No. It is a broad, single-pass review and may miss issues. The skill itself points to the security-reviewer skill for deeper security work, and a dedicated audit or penetration test is still the right choice for serious assurance.
Why does the review ask me questions about my change?
The skill begins by summarizing the intent of the change. If it cannot do that from the description, it asks the author to clarify, and unexplained changes give weaker results.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you want consistent, prioritized feedback on pull requests and quality audits, and you can give the agent context about each change. Pick a dedicated security audit or penetration test for serious security assurance, and keep a human in charge of merge decisions.
Download the skill