🚀 PR-Sentinel

An AI-powered code review application that integrates with GitHub pull requests to automatically analyze code changes using Google Gemini API.

---

📌 Features

• Connects to GitHub via Personal Access Token (PAT).
• Listens to pull request events (via GitHub Actions workflow).
• Fetches the diff/patch of PR changes.
• Sends the diff to Google Gemini API for AI-powered review.
• Prints AI suggestions for better code quality.

---

🛠️ Setup Instructions

1. Create GitHub Repository

• Initialize a new GitHub repo (e.g., AI-code-review-demo).
• Add a sample main.py file (or any code file) to test with.
• Clone the repo locally:

  git clone https://github.com/<your-username>/AI-code-review-demo.git
  cd AI-code-review-demo `
  

---

2. Create a GitHub Personal Access Token (PAT)

1. Go to your GitHub account:
   Profile > Settings > Developer Settings > Personal Access Tokens > Fine-grained Tokens

2. Click Generate new token.

3. Select "Only select repositories" and choose the repository where you want to enable AI code review.
   (This ensures the token is restricted to just your project and not all repos.)

4. Under Repository permissions, grant the following:

   • Contents → Read and Write (needed to fetch and comment on PRs)
   • Pull requests → Read and Write (to post AI review comments on PRs)
   • Workflows → Read and Write (so Actions can use the token)

5. (Optional) Under Account permissions, you can leave defaults unless your repo setup requires more.

6. Set an expiry date for security (recommended).

7. Generate the token and copy it immediately — GitHub will only show it once.

8. Save this token in your .env file as:

   GITHUB_PAT_TOKEN=ghp_yourGeneratedTokenHere
   

---

3. Add Google Gemini API Key

1. Go to Google AI Studio.

2. Create an API key.

3. Save it in .env file.

---

4. Create .env File

In the project root, create a .env file:

GITHUB_PAT_TOKEN=ghp_yourgithubtoken GEMINI_API_KEY=AIzaSyDxxxxxxx

---

Note:

To write an automated comment on the PR use this command on git bash:

curl -X POST \
  -H "Authorization: token YOUR_GITHUB_TOKEN" \
  -H "Accept: application/vnd.github+json" \
  https://api.github.com/repos/OWNER/REPO/issues/PR_NUMBER/comments \
  -d '{"body":"🤖 AI Code Review Suggestion:\n\nYour review text goes here"}'

5. Install Dependencies

Inside your project folder:

pip install requests python-dotenv google-generativeai

---

6. Project Structure

│── AI_review/
│   ├── review_pr.py   # Main script to fetch PR diff and send to Gemini
│   ├── script.py      # Additional utilities
│── .env               # Environment variables
│── .gitignore
│── README.md
│── main.py            # Sample file

---

7. GitHub Actions Workflow

Inside .github/workflows/code_review.yml:

name: AI Code Review

on:
  pull_request:
    types: [opened, synchronize]

jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout repo
        uses: actions/checkout@v3

      - name: Set up Python
        uses: actions/setup-python@v4
        with:
          python-version: "3.10"

      - name: Install dependencies
        run: pip install requests python-dotenv google-generativeai

      - name: Run AI Code Review
        env:
          GITHUB_PAT_TOKEN: ${{ secrets.GITHUB_PAT_TOKEN }}
          GEMINI_API_KEY: ${{ secrets.GEMINI_API_KEY }}
        run: python AI_review/review_pr.py

8. Running the Script Locally

To test without workflow:

python AI_review/review_pr.py

It will:

• Fetch latest PR diff.

• Send to Gemini.

• Print AI review in console.

---

✅ Example Output

{ "status": "success", "pull_request": 12, "ai_review": "Consider adding error handling for file operations in main.py..." }

---

🔑 Notes

• If you used a dummy webhook earlier, you don't need it now since GitHub Actions workflow handles PR events.

• Always keep .env and tokens secure (never commit them).

• For merging branches:

  git checkout main git merge branch-2 git push origin main

---

🎯 Summary

This project demonstrates end-to-end AI integration with GitHub PRs, automating code reviews using Gemini API.