Integrate Codegen with GitHub Actions to automate code reviews, issue triage, and other development workflows. This guide shows you how to use the Codegen Python SDK within GitHub Actions workflows.

This approach is necessary because GitHub webhooks don’t fire for bot users, preventing direct @codegen mentions in automated workflows. Using GitHub Actions with the Codegen API provides a reliable alternative.

​
Prerequisites

Before setting up GitHub Actions with Codegen, ensure you have:

  • A GitHub repository with Actions enabled
  • A Codegen account with API access
  • Your organization ID and API token from codegen.com/developer
  • The Codegen GitHub App installed on your repository

​
Basic Setup

​
1. Store Your Credentials

Add your Codegen credentials as GitHub repository secrets:

  1. Go to your repository → Settings → Secrets and variables → Actions
  2. Add the following secrets:
    • CODEGEN_ORG_ID: Your organization ID from codegen.com/developer
    • CODEGEN_API_TOKEN: Your API token from codegen.com/developer

​
2. Create a GitHub Actions Workflow

Create .github/workflows/codegen-review.yml in your repository:

name: Codegen PR Review

on:
  pull_request:
    types: [opened, synchronize]
    branches: [main, develop]

jobs:
  codegen-review:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout code
        uses: actions/checkout@v4

      - name: Set up Python
        uses: actions/setup-python@v4
        with:
          python-version: '3.11'

      - name: Install Codegen SDK
        run: |
          pip install codegen

      - name: Run Codegen Review
        env:
          CODEGEN_ORG_ID: ${{ secrets.CODEGEN_ORG_ID }}
          CODEGEN_API_TOKEN: ${{ secrets.CODEGEN_API_TOKEN }}
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          python .github/scripts/codegen_review.py

​
3. Create the Review Script

Create .github/scripts/codegen_review.py:

import os
import sys
from codegen.agents.agent import Agent

def main():
    # Get environment variables
    org_id = os.getenv('CODEGEN_ORG_ID')
    api_token = os.getenv('CODEGEN_API_TOKEN')
    github_token = os.getenv('GITHUB_TOKEN')
    
    if not all([org_id, api_token]):
        print("Error: Missing required environment variables")
        sys.exit(1)
    
    # Get PR information from GitHub context
    pr_number = os.getenv('GITHUB_EVENT_PATH')
    repo_name = os.getenv('GITHUB_REPOSITORY')
    
    # Initialize Codegen agent
    agent = Agent(
        org_id=org_id,
        token=api_token
    )
    
    # Create review prompt
    prompt = f"""
    Please review PR #{pr_number} in repository {repo_name}.
    
    Focus on:
    - Code quality and best practices
    - Potential bugs or security issues
    - Performance considerations
    - Documentation and testing
    
    Provide specific, actionable feedback with line-by-line comments where appropriate.
    """
    
    try:
        # Run the agent
        task = agent.run(prompt=prompt)
        
        # Wait for completion (with timeout)
        max_attempts = 30  # 5 minutes with 10-second intervals
        attempt = 0
        
        while attempt < max_attempts:
            task.refresh()
            print(f"Task status: {task.status}")
            
            if task.status == "completed":
                print("✅ Review completed successfully!")
                if hasattr(task, 'result') and task.result:
                    print(f"Result: {task.result}")
                break
            elif task.status == "failed":
                print("❌ Review failed")
                sys.exit(1)
            
            attempt += 1
            time.sleep(10)
        
        if attempt >= max_attempts:
            print("⏰ Review timed out")
            sys.exit(1)
            
    except Exception as e:
        print(f"Error running Codegen review: {e}")
        sys.exit(1)

if __name__ == "__main__":
    main()

​
Advanced Configuration

​
Conditional Reviews

Only run reviews for specific conditions:

name: Conditional Codegen Review

on:
  pull_request:
    types: [opened, synchronize]

jobs:
  check-conditions:
    runs-on: ubuntu-latest
    outputs:
      should-review: ${{ steps.check.outputs.should-review }}
    steps:
      - name: Check if review needed
        id: check
        run: |
          # Skip if PR is from dependabot
          if [[ "${{ github.actor }}" == "dependabot[bot]" ]]; then
            echo "should-review=false" >> $GITHUB_OUTPUT
          # Skip if PR is a draft
          elif [[ "${{ github.event.pull_request.draft }}" == "true" ]]; then
            echo "should-review=false" >> $GITHUB_OUTPUT
          # Skip if PR has "skip-review" label
          elif [[ "${{ contains(github.event.pull_request.labels.*.name, 'skip-review') }}" == "true" ]]; then
            echo "should-review=false" >> $GITHUB_OUTPUT
          else
            echo "should-review=true" >> $GITHUB_OUTPUT
          fi

  codegen-review:
    needs: check-conditions
    if: needs.check-conditions.outputs.should-review == 'true'
    runs-on: ubuntu-latest
    # ... rest of the job

​
Custom Review Types

Create different review workflows for different scenarios:

# .github/scripts/codegen_security_review.py
def security_review():
    prompt = f"""
    Perform a security-focused review of PR #{pr_number} in {repo_name}.
    
    Look specifically for:
    - SQL injection vulnerabilities
    - XSS vulnerabilities
    - Authentication/authorization issues
    - Sensitive data exposure
    - Insecure dependencies
    
    Provide detailed security recommendations.
    """
    # ... rest of implementation

# .github/scripts/codegen_performance_review.py
def performance_review():
    prompt = f"""
    Analyze PR #{pr_number} in {repo_name} for performance implications.
    
    Focus on:
    - Database query efficiency
    - Memory usage patterns
    - CPU-intensive operations
    - Caching opportunities
    - Scalability concerns
    
    Suggest specific optimizations.
    """
    # ... rest of implementation

​
Error Handling and Retries

Implement robust error handling:

import time
import logging
from typing import Optional

def run_with_retry(agent: Agent, prompt: str, max_retries: int = 3) -> Optional[str]:
    """Run agent with retry logic and proper error handling."""
    
    for attempt in range(max_retries):
        try:
            task = agent.run(prompt=prompt)
            
            # Wait for completion with exponential backoff
            max_wait_time = 300  # 5 minutes
            check_interval = 10
            total_wait = 0
            
            while total_wait < max_wait_time:
                task.refresh()
                logging.info(f"Attempt {attempt + 1}: Task status: {task.status}")
                
                if task.status == "completed":
                    return task.result if hasattr(task, 'result') else "Review completed"
                elif task.status == "failed":
                    raise Exception(f"Task failed: {getattr(task, 'error', 'Unknown error')}")
                
                time.sleep(check_interval)
                total_wait += check_interval
            
            raise Exception("Task timed out")
            
        except Exception as e:
            logging.error(f"Attempt {attempt + 1} failed: {e}")
            if attempt == max_retries - 1:
                raise
            time.sleep(2 ** attempt)  # Exponential backoff
    
    return None

​
Environment-Specific Workflows

​
Development Environment

# .github/workflows/codegen-dev-review.yml
name: Codegen Dev Review

on:
  pull_request:
    types: [opened, synchronize]
    branches: [develop, feature/*]

jobs:
  dev-review:
    runs-on: ubuntu-latest
    steps:
      # ... setup steps
      - name: Run Development Review
        env:
          CODEGEN_ORG_ID: ${{ secrets.CODEGEN_ORG_ID }}
          CODEGEN_API_TOKEN: ${{ secrets.CODEGEN_API_TOKEN }}
          REVIEW_TYPE: "development"
        run: |
          python .github/scripts/codegen_review.py --type development

​
Production Environment

# .github/workflows/codegen-prod-review.yml
name: Codegen Production Review

on:
  pull_request:
    types: [opened, synchronize]
    branches: [main, master]

jobs:
  prod-review:
    runs-on: ubuntu-latest
    steps:
      # ... setup steps
      - name: Run Production Review
        env:
          CODEGEN_ORG_ID: ${{ secrets.CODEGEN_ORG_ID }}
          CODEGEN_API_TOKEN: ${{ secrets.CODEGEN_API_TOKEN }}
          REVIEW_TYPE: "production"
        run: |
          python .github/scripts/codegen_review.py --type production --strict

​
Best Practices

​
Security

  • Never commit API tokens: Always use GitHub Secrets
  • Use least privilege: Create dedicated API tokens with minimal required permissions
  • Rotate tokens regularly: Update API tokens periodically
  • Monitor usage: Track API usage to detect anomalies

​
Performance

  • Set appropriate timeouts: Don’t let workflows run indefinitely
  • Use caching: Cache Python dependencies to speed up workflow execution
  • Parallel execution: Run multiple review types in parallel when possible
  • Rate limiting: Respect API rate limits to avoid throttling

​
Reliability

  • Implement retries: Handle transient failures gracefully
  • Add logging: Include detailed logging for debugging
  • Fail gracefully: Don’t block PRs if reviews fail
  • Monitor workflows: Set up alerts for workflow failures

​
Troubleshooting

​
Common Issues

Issue: “No repos found in the organization”

Solution: Verify your organization ID is correct and the repository 
is accessible to your Codegen organization.

Issue: “Authentication failed”

Solution: Check that your API token is valid and properly set in 
GitHub Secrets. Tokens may expire and need renewal.

Issue: “Task timeout”

Solution: Increase the timeout duration or break large PRs into 
smaller chunks. Complex reviews may take longer to complete.

Issue: “Workflow not triggering”

Solution: Check the workflow triggers and ensure the PR meets 
the specified conditions (branch names, PR types, etc.).

​
Debugging

Enable debug logging in your workflow:

- name: Run Codegen Review
  env:
    CODEGEN_DEBUG: "true"
    CODEGEN_LOG_LEVEL: "DEBUG"
  run: |
    python .github/scripts/codegen_review.py

Add debug output to your Python script:

import logging

logging.basicConfig(
    level=logging.DEBUG if os.getenv('CODEGEN_DEBUG') else logging.INFO,
    format='%(asctime)s - %(levelname)s - %(message)s'
)

logger = logging.getLogger(__name__)

​
Examples

​
Complete Working Example

Here’s a complete, production-ready workflow:

View Complete Example

See a full working example with all files and configurations in our examples repository.

​
Integration with Existing CI/CD

Integrate Codegen reviews with your existing CI/CD pipeline:

name: CI/CD with Codegen

on:
  pull_request:
    types: [opened, synchronize]

jobs:
  tests:
    runs-on: ubuntu-latest
    steps:
      - name: Run tests
        run: npm test

  codegen-review:
    needs: tests  # Only run review if tests pass
    if: success()
    runs-on: ubuntu-latest
    steps:
      # ... Codegen review steps

  deploy:
    needs: [tests, codegen-review]
    if: success() && github.ref == 'refs/heads/main'
    runs-on: ubuntu-latest
    steps:
      - name: Deploy
        run: npm run deploy

​
Next Steps