LangGraph Code Review Agent

This repository implements a LangGraph agent that performs a code review on a Python function. The graph consists of three nodes:

1. draft_review – generates an initial review.
2. reflect – a critic that scores the review on four criteria (PEP8, type hints, edge cases, naming) and decides whether a rewrite is required.
3. rewrite – rewrites the weakest part of the review.

The graph runs for a maximum of max_rounds (default 2). The demo can be executed with python -m main.

Requirements

langgraph>=0.2.0
langchain-openai>=0.2.0
python-dotenv>=1.0.0

Usage

# Install dependencies
pip install -r requirements.txt

# Run the demo
python -m main

The demo uses a simple sort_numbers function. The output shows the final review, the scores for each criterion, the verdict, and the number of rounds performed.

Environment Variables

The agent uses OpenAI. Set the OPENAI_API_KEY environment variable before running.