Lena Code Review Assistant

Accelerate Your Code Reviews with AI

Transform your development workflow with intelligent, AI-powered code reviews. By integrating OpenAI’s language models directly into your Azure DevOps pipeline, this extension elevates code quality by delivering faster, more insightful reviews.

Key Benefits

• Automated Reviews
  Allow AI to scrutinize your code changes, pinpoint bugs, detect performance bottlenecks, and recommend industry best practices.

• Intelligent Insights
  Leverage advanced natural language processing to receive comprehensive, actionable feedback on every pull request.

• Accelerated Process
  Reduce manual review time, enabling your team to focus on mission-critical development tasks.

• Fully Customizable
  Tailor the AI model settings, specify file inclusions/exclusions, and add custom prompts to align with your project’s coding standards.

Prerequisites

• An OpenAI API Key
• Important: Ensure that Allow Scripts to Access OAuth Token is enabled in your pipeline. For more information, see this guide.

Getting Started

Integrate Lena Code Review Assistant into your Azure DevOps pipeline with the following configuration:

trigger:
  branches:
    exclude:
      - '*'

pr:
  branches:
    include:
      - '*'

jobs:
- job: CodeReview
  pool:
    vmImage: 'ubuntu-latest'
  steps:
  - task: LenaCodeReviewAssistant@1
    inputs:
      api_key: $(OpenAI_ApiKey)
      ai_model: 'gpt-4o'
      file_extensions: ''
      file_excludes: ''
      prompt: 'You are a senior developer specializing in .NET C# development. Your task is to review the code changes provided in a Pull Request (in unidiff format) according to the following guidelines:\n- Use bullet points if you have multiple comments.\n- Clearly highlight any bugs you find.\n- Identify and emphasize any major performance issues.\n- Detail any missed opportunities for applying best practices.\n- Do not mention minor issues such as indentation, styling, or other trivial nitpicks.\n- Provide clear, concise feedback.\n- Include code examples where possible to illustrate the issue.\n- Only offer instructions for improvements.\n- If there are no improvement instructions to provide, respond with only \"NO_COMMENT\".\nThe response should be formatted in Markdown.'