Scenario: Automated Code Review

Learn how to leverage LibrAgent for analyzing pull requests, detecting anti-patterns, reviewing diffs, and suggesting refactoring.

---

🎯 Objectives

• Perform automated quality checks on local code diffs or commit ranges.
• Detect security vulnerabilities, code smells, and performance bottlenecks.
• Generate structured code review summaries with concrete refactoring snippets.

---

📋 Step-by-Step Guide

Step 1: Specify Target Code & Scope

Prompt the agent with your target files, branches, or PR diffs:

Review the uncommitted changes in the src/components/ directory and highlight potential memory leaks or missing React dependency arrays.

Step 2: Set Review Criteria

You can instruct the agent on specific coding conventions or guidelines:

Analyze the changes against our clean code criteria:
- Single Responsibility Principle (SRP)
- Proper TypeScript type safety (no explicit `any`)
- Error handling completeness

Step 3: Inspect Review Report

The agent will execute tools (e.g. workspace__readFile, workspace__runShell) to review the code and output a structured review report:

## 🔍 Code Review Summary

### ⚠️ Recommendations

1. **`useSettingsPageController.ts`**: Missing cleanup function in `useEffect`.
2. **`AIModelsTab.tsx`**: Implicit `any` type found on model configuration callback.

### 💡 Refactoring Snippet

...