Installs into .claude/skills of the current project.
Are you the author of Templates?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/notysoty-templates-openagentskills)
---
name: Your Code Review Skill Name
description: One sentence describing what this review skill checks for and fixes.
category: coding
difficulty: beginner
tags:
- code-review
- your-tag
author: your-github-username
---
# Your Code Review Skill Name
## What this skill does
This skill directs the agent to review code for [specific concern β e.g., security issues, performance bottlenecks, accessibility violations]. It produces a prioritized list of findings with severity levels and concrete fix suggestions.
Use this when you want a focused review on [concern] rather than a general code review.
## How to use
### Claude Code / Cline
Copy this file to `.agents/skills/your-skill-name/SKILL.md` in your project root.
Then ask:
- *"Review `src/auth.ts` using the [Skill Name] skill."*
- *"Run the [Skill Name] skill on all files in `src/api/`."*
### Cursor
Add the **The Prompt / Instructions** section below to your `.cursorrules` file, or paste it into a Cursor chat before making your request.
### Codex
Paste the instructions section into the chat along with the file(s) you want reviewed.
## The Prompt / Instructions for the Agent
When asked to review code using this skill, follow these steps:
1. **Read the file(s).** Understand the structure, purpose, and any existing patterns before flagging issues.
2. **Check for [concern #1].** Look for [specific pattern or anti-pattern]. Flag any instance where [condition].
3. **Check for [concern #2].** Look for [specific pattern]. Flag cases where [condition].
4. **Check for [concern #3].** Look for [specific pattern]. Flag any [condition].
5. **Prioritize findings.** Sort issues into:
- π΄ **Critical** β must fix before merging (security risk, data loss, crash)
- π **High** β should fix soon (performance, correctness)
- π‘ **Medium** β fix when convenient (maintainability, readability)
- π’ **Low** β optional improvement (style, minor cleanup)
6. **For each finding, output:**
- File and line number
- Severity level
- What the problem is
- Why it matters
- A concrete fix with a code snippet
7. **Summarize at the end:** Total issues found by severity, and an overall assessment (Ready to merge / Needs minor fixes / Needs major fixes).
## Example
**Input to Agent:**
> "Review `src/middleware/auth.ts` using the [Skill Name] skill."
**Output from Agent:**
```
## Code Review: src/middleware/auth.ts
### π΄ Critical
**Line 42 β [Issue title]**
Problem: [Description of what's wrong]
Why it matters: [Impact if not fixed]
Fix:
```ts
// Before
const token = req.headers.authorization;
// After
const token = req.headers.authorization?.replace('Bearer ', '');
if (!token) return res.status(401).json({ error: 'Missing token' });
```
### π‘ Medium
**Line 78 β [Issue title]**
Problem: [Description]
Fix: [Suggested change]
---
Summary: 1 critical, 0 high, 1 medium, 0 low. Needs fixes before merge.
```
## Notes
- This skill focuses on [concern] only. For a general review, use the Code Review Checklist skill.
- Works best on files under ~500 lines. For larger files, run it on specific functions or sections.
- [Any other known limitations or tips]