Guide · Agent skill
How to use the Code Review and Quality agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The code-review-and-quality skill gives an AI agent a structured checklist for reviewing code changes before they are merged. It is written for developers who want consistent reviews of their own code, a teammate's code, or code produced by another agent. After reading this guide you will know what the skill does, how to install it, how to ask for a review in plain language, and what its limits are.
Code Review and Qualityaddyosmani/agent-skills · MIT Download the skill Quick start
- 1Download the ZIP. Download the code-review-and-quality ZIP from this skill's page on the directory.
- 2Unzip the skill. Unzip the file so that SKILL.md sits in a code-review-and-quality folder under ~/.claude/skills/ or the project's .claude/skills/.
- 3Or upload it in the apps. In the Claude apps, open Settings → Capabilities → Skills and upload the ZIP instead.
- 4Ask for a review. Paste a diff or refer to a pull request and say you want it reviewed before merging.
- 5Read the five-axis feedback. Check the agent's comments on correctness, readability, architecture, security, and performance.
- 6Apply or question the fixes. Evaluate the proposed restructurings and decide which to apply, then confirm with a human reviewer for critical code.
What the Code Review and Quality skill makes the agent do
When installed, the agent reviews code changes such as diffs, pull requests, or code pasted into chat. It evaluates each change against five areas: correctness, readability and simplicity, architecture, security, and performance.
The skill also checks that a change is a reasonable size and suggests specific restructurings rather than only pointing out problems. Its approval standard is to approve a change when it clearly improves overall code health, even if it is not perfect.
- Correctness: does the code do what it claims, including edge cases and error paths
- Readability and simplicity: can another engineer follow it without the author explaining
- Architecture: does it fit existing patterns and module boundaries
- Security: input validation, secrets, authorization, and untrusted external data
- Performance: N+1 queries, unbounded loops, missing pagination, and similar issues
- Change sizing: flags changes that are too large and suggests how to split them
How to install the skill
Download the ZIP file from the skill's page on this directory. The skill's package contains a single file, SKILL.md. Unzip it into one of the following locations, depending on whether you want the skill for all projects or one project.
For a personal install available to every project, unzip it into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder. If you use the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside your project
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
How to trigger the skill with natural requests
You do not need a special command. The skill's description tells the agent to use it before merging any change, when reviewing code written by yourself, another agent, or a human, and when asked to review a diff or pull request, even when the diff is pasted inline.
Ask in plain language and say what you want reviewed. Including the diff, the pull request reference, or the files involved helps the agent work from the actual code.
- Review this diff before I merge it: [paste diff]
- Review the changes in this pull request for security and performance problems
- Another agent wrote this code. Check it before it goes into main.
- I fixed this bug. Review both the fix and the regression test.
Example: a review from start to finish
Suppose you paste a diff that adds a new conditional into an existing checkout flow and ask for a review. The skill directs the agent to assess correctness first, then readability, architecture, security, and performance.
If the agent finds a structural problem, the skill asks it to propose a named fix, such as replacing a chain of conditionals with a dispatcher or extracting a helper, rather than only saying the code is complex. Review comments therefore tend to arrive with a suggested change you can evaluate.
What the skill needs and its safety notes
The skill lists no requirements. It does not run scripts and does not use the network. The file is written as guidance for the agent's reasoning and does not itself run any tools, so it cannot test your code or verify its claims.
Because the review is guidance only, treat it as a second reader rather than a final authority. Keep human sign-off on security-critical or production code.
- Requires: nothing listed
- Runs scripts: no
- Uses network: no
- The agent's review can miss issues, and the skill provides no tool-based verification
Tips for getting better reviews
Give the agent a small, focused change. The skill's sizing guidance says changes around 100 lines are easy to review, around 300 is acceptable for one logical change, and around 1000 is too large and should be split.
Tell the agent about the project's conventions and any spec or task the change should meet, since the correctness and readability checks depend on them.
- Split large changes by stack, file group, horizontal layer, or vertical slice
- Separate refactoring from feature work into different changes
- Ask the agent to check that the commit message's first line is short, imperative, and standalone
- Ask the agent to ask whether a large file should be split before adding more code to it
What people use it for
- Reviewing your own pull request before asking a teammate to look at it
- Checking code that another AI agent or model produced before it enters the main branch
- Reviewing a teammate's diff with a consistent checklist
- Getting review comments that propose concrete restructurings, not only problems
- Reviewing a bug fix together with its regression test
Frequently asked questions
How do I install the code-review-and-quality skill?
Download the ZIP from this skill's page and unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder. In the Claude apps, you can upload the ZIP under Settings → Capabilities → Skills.
Does this skill cost money or what is its licence?
The skill is listed under the MIT licence. This guide does not state any price, so check the directory page for current terms.
What does the skill require to run?
The skill lists no requirements. It runs no scripts and uses no network access. It is a single SKILL.md file of guidance for the agent.
Can the skill replace a human code review?
No. The skill gives guidance only and has no tool-based verification, and the agent's review can miss issues. Keep human sign-off for security-critical or production code.
Do I need to be a programmer to use it?
The skill assumes familiarity with software engineering terms. Non-programmers who need code explained from scratch should look for a different resource.
How is this different from other code review tools?
This skill is a checklist that shapes how the agent reasons about a change. It does not run linters, tests, or scanners, so it is best combined with your existing tooling and human reviewers.
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 a consistent five-axis checklist and review comments that propose concrete fixes, especially for code written by AI agents or teammates. Pick something else, or add human review, when the code is security-critical or production-facing, or if you need automated tool-based checks, since this skill only guides the agent's reasoning.
Download the skill