Code Review

Overview

Code Review brings the rigor of a senior engineer’s review to every pull request. Instead of waiting hours (or days) for human reviewers, this skill analyzes your code changes in seconds — catching bugs, security vulnerabilities, and performance issues that slip past linters and type checkers.

Unlike simple linting tools, Code Review understands the semantic context of your changes. It knows when a variable rename breaks an API contract, when a database query introduces an N+1 problem, or when an error handler silently swallows critical exceptions.

The skill works with any programming language and integrates directly into your existing code review workflow on GitHub, GitLab, or Bitbucket.

How It Works

1. Trigger — The skill activates when a new pull request is opened or updated
2. Analyze — It parses the diff, understands the context of surrounding code, and identifies issues across multiple severity levels
3. Review — Inline comments are posted directly on the PR with explanations, severity tags (🔴 Blocker / 🟡 Suggestion / 💭 Nit), and suggested fixes
4. Learn — Over time, it learns your team’s patterns and preferences to reduce noise

Use Cases

• Pre-merge quality gate — Catch bugs before they reach main branch
• Security scanning — Detect SQL injection, XSS, hardcoded secrets, and other OWASP vulnerabilities
• Onboarding — Help junior developers learn best practices through review feedback
• Consistency — Enforce coding standards across large teams
• Performance — Identify N+1 queries, memory leaks, and unnecessary re-renders

Getting Started

# Install the GitHub App or add the webhook
npm install @ai-review/cli

# Configure for your repo
ai-review init --provider github --repo your-org/your-repo

# Run manually on a PR
ai-review check --pr 42

Example

Given a pull request with this change:

def get_users(db, name):
    query = f"SELECT * FROM users WHERE name = '{name}'"
    return db.execute(query)

The skill would flag:

🔴 Security: SQL Injection Risk — User input name is interpolated directly into the query string. Use parameterized queries: db.execute("SELECT * FROM users WHERE name = %s", (name,))

Alternatives

• CodeRabbit — AI-powered code review with GitHub integration
• Sourcery — Python-focused automated code review
• Amazon CodeGuru — AWS-native code review service