expert-code-reviewer

umaranit's avatarfrom umaranit

Performs a deep-dive code review based on industry standards and local project rules. Trigger this when a user asks to "review", "audit", or "check" code.

⭐0stars🔀0forks📁View on GitHub🕐Updated Jan 8, 2026
[Code Review]

When & Why to Use This Skill

The Expert Code Reviewer skill automates comprehensive code audits by evaluating source code against industry best practices and custom project-specific rules. It streamlines the software development lifecycle by identifying security vulnerabilities, performance bottlenecks, and maintainability issues, providing structured, actionable feedback through a prioritized traffic-light reporting system.

Use Cases

  • Automated Security Auditing: Detect hardcoded secrets, SQL injection risks, and insecure dependencies to prevent vulnerabilities before they reach production.
  • Enforcing Team Coding Standards: Automatically verify compliance with local project rules and 'Modern Clean Code' standards defined in configuration files.
  • Performance Optimization: Identify inefficient loops and heavy computations within render cycles to improve application speed and resource management.
  • Technical Debt Reduction: Analyze function length and variable naming conventions to ensure long-term code maintainability and readability.
  • Pre-Merge Quality Checks: Use the structured 'Traffic Light' feedback to quickly prioritize critical fixes and architectural improvements during the Pull Request process.
nameexpert-code-reviewer
descriptionPerforms a deep-dive code review based on industry standards and local project rules. Trigger this when a user asks to "review", "audit", or "check" code.

Expert Code Reviewer Skill

1. Context Discovery (Mandatory First Step)

Before providing feedback, you MUST:

  1. Check the root directory for a skill-config.json file.
  2. If found, identify the coding_standard and forbidden_patterns defined there.
  3. If not found, use the default "Modern Clean Code" standards.

2. Review Process

Follow these steps in order:

  • Security Check: Look for hardcoded secrets, SQL injection risks, and insecure dependencies.
  • Maintainability: Ensure functions are under 30 lines and variables are descriptively named.
  • Performance: Identify unnecessary loops or heavy computations inside render cycles.
  • Project Alignment: Verify the code follows the directory structure defined in the project's README.md.

3. Output Format

Always present your review in a "Traffic Light" format:

  • 🔴 Critical: Fix immediately (Security/Bugs).
  • 🟡 Warning: Architectural improvements.
  • 🟢 Praise: Point out what was done well.