GitLab AI Assistant

AI-powered GitLab integration for VS Code that automates branch creation, commits, merge request management, and provides intelligent code review.

Features

• 🌿 AI Branch Creation: Create branches with intelligent naming based on your description
• 💬 Smart Commits: Generate conventional commit messages from your changes
• 🚀 Automated MR Creation: Push and create merge requests with AI-generated descriptions
• 🔍 Intelligent Code Review: Get comprehensive code reviews using AI (syntax, logic, business context)
• ⚡ Full Workflow Automation: Complete workflow from branch creation to MR review
• 🎯 Multiple LLM Support: Works with OpenAI, Anthropic, Ollama, or custom endpoints
• 📋 Customizable Conventions: Configure branch naming, commit messages, and review modes

Requirements

• VS Code 1.85.0 or higher
• GitLab account (self-hosted or GitLab.com)
• GitLab Personal Access Token with api scope
• LLM API access (OpenAI, Anthropic, Ollama, or custom)

Installation

From Source

1. Clone this repository:

   git clone https://github.com/CHashtager/gitlab-ai-assistant
   cd gitlab-ai-extension
   

2. Install dependencies:

   npm install
   

3. Compile the extension:

   npm run compile
   

4. Package the extension:

   npm run package
   

5. Install the generated .vsix file in VS Code:

   • Open VS Code
   • Go to Extensions view (Cmd+Shift+X)
   • Click "..." menu → "Install from VSIX..."
   • Select the generated .vsix file

Configuration

Quick Setup

1. After installation, run the command: GitLab AI: Configure Settings
2. Follow the interactive setup wizard to configure:
   • GitLab connection (URL and token)
   • LLM provider and credentials
   • Branch naming conventions
   • Commit message formats
   • Code review preferences

Manual Configuration

Open VS Code settings (Cmd+,) and search for "GitLab AI" or edit .vscode/settings.json:

{
  "gitlabAI.gitlabUrl": "https://gitlab.yourcompany.com",
  "gitlabAI.gitlabToken": "your-gitlab-token",
  "gitlabAI.llmProvider": "openai",
  "gitlabAI.llmApiKey": "your-api-key",
  "gitlabAI.llmModel": "gpt-4",
  "gitlabAI.branchNamingConvention": "{type}/{ticket}-{description}",
  "gitlabAI.commitMessageConvention": "conventional",
  "gitlabAI.reviewMode": "comprehensive",
  "gitlabAI.defaultTargetBranch": "main",
  "gitlabAI.enableDraftMR": true
}

Configuration Options

GitLab Settings

• gitlabAI.gitlabUrl: Your GitLab instance URL (e.g., https://gitlab.com or self-hosted)
• gitlabAI.gitlabToken: Personal Access Token with api scope

LLM Settings

• gitlabAI.llmProvider: Choose from openai, anthropic, ollama, or custom
• gitlabAI.llmApiUrl: API URL for custom/Ollama endpoints
• gitlabAI.llmApiKey: API key for your LLM provider
• gitlabAI.llmModel: Model name (e.g., gpt-4, claude-3-opus-20240229, llama2)

Branch Settings

• gitlabAI.branchNamingConvention: Template with placeholders: {type}, {ticket}, {description}, {username}, {date}
• gitlabAI.branchTypes: Array of allowed branch types (e.g., ["feature", "bugfix", "hotfix"])

Commit Settings

• gitlabAI.commitMessageConvention: conventional, angular, gitmoji, or custom
• gitlabAI.customCommitTemplate: Custom template when using custom convention

Review Settings

• gitlabAI.reviewMode: syntax, logic, business, or comprehensive
• gitlabAI.businessContextFile: Path to business context file (default: .gitlab-ai-context.md)

Merge Request Settings

• gitlabAI.defaultTargetBranch: Default target branch for MRs (default: main)
• gitlabAI.enableDraftMR: Create MRs as draft by default
• gitlabAI.autoAssignReviewers: Auto-assign reviewers based on CODEOWNERS

Usage

Commands

Access commands via Command Palette (Cmd+Shift+P):

• GitLab AI: Create New Branch - Create a branch with AI-generated name
• GitLab AI: Commit Changes with AI Message - Generate and commit with AI message
• GitLab AI: Push & Create Merge Request - Push and create MR with AI description
• GitLab AI: Review Current Changes - AI review of local changes
• GitLab AI: Review Merge Request - Review an MR with inline comments
• GitLab AI: Full Workflow - Complete automation (Branch → Commit → Push → MR → Review)
• GitLab AI: Configure Settings - Interactive configuration wizard

Keyboard Shortcuts

• Cmd+Shift+G C - Commit Changes
• Cmd+Shift+G M - Push & Create MR
• Cmd+Shift+G F - Full Workflow

Status Bar

Click the "GitLab AI" icon in the status bar for quick access to all commands.

Development

Building from Source

# Install dependencies
npm install

# Compile TypeScript
npm run compile

# Watch mode for development
npm run watch

# Lint code
npm run lint

# Run tests
npm test

# Package extension
npm run package

Project Structure

src/
  commands/        # Command implementations
  services/        # Core services (Git, GitLab, LLM)
  utils/          # Utilities and configuration
  types/          # TypeScript type definitions
  extension.ts    # Extension entry point

Testing the Extension

1. Test Configuration

Run the configuration test:

Command Palette → GitLab AI: Configure Settings → Test Configuration

This will verify:

• GitLab connection and authentication
• LLM API connectivity
• Configuration validity

2. Test Individual Features

Create Branch

1. Open a Git repository
2. Run: GitLab AI: Create New Branch
3. Enter a description (e.g., "add user authentication")
4. Select branch type
5. Verify branch is created locally and on GitLab

Commit Changes

1. Make some code changes
2. Run: GitLab AI: Commit Changes with AI Message
3. Review the AI-generated commit message
4. Confirm to commit

Create Merge Request

1. Make changes and commit
2. Run: GitLab AI: Push & Create Merge Request
3. Verify MR is created on GitLab with AI-generated description

Code Review

1. Make some changes (don't commit yet)
2. Run: GitLab AI: Review Current Changes
3. Check the Problems panel for AI-generated review comments

3. Test Full Workflow

1. Run: GitLab AI: Full Workflow
2. Follow the prompts for each step
3. Verify the complete flow:
   • Branch created
   • Changes committed
   • Pushed to GitLab
   • MR created
   • Code review performed

4. Testing with Different LLM Providers

OpenAI

{
  "gitlabAI.llmProvider": "openai",
  "gitlabAI.llmApiKey": "sk-...",
  "gitlabAI.llmModel": "gpt-4"
}

Anthropic

{
  "gitlabAI.llmProvider": "anthropic",
  "gitlabAI.llmApiKey": "sk-ant-...",
  "gitlabAI.llmModel": "claude-3-opus-20240229"
}

Ollama (Local)

# Start Ollama
ollama serve

# Pull a model
ollama pull llama2

{
  "gitlabAI.llmProvider": "ollama",
  "gitlabAI.llmApiUrl": "http://localhost:11434",
  "gitlabAI.llmModel": "llama2"
}

Business Context

Create a .gitlab-ai-context.md file in your project root to provide business context for code reviews:

# Project Context for AI Code Review

## Project Description
E-commerce platform for selling digital products with subscription support.

## Architecture
- Frontend: React + TypeScript
- Backend: Node.js + Express
- Database: PostgreSQL
- Payment: Stripe integration

## Coding Standards
- Use TypeScript strict mode
- Follow Airbnb style guide
- All API endpoints must have authentication
- Use async/await, not callbacks

## Domain Terms
- **Product**: Digital item for sale
- **Subscription**: Recurring payment plan
- **Credit**: In-app currency

## Security Requirements
- All user inputs must be validated
- No sensitive data in logs
- Rate limiting on all API endpoints

## Performance Requirements
- API response time < 200ms
- Support 1000 concurrent users

Troubleshooting

Extension Not Activating

• Check VS Code version (must be ≥1.85.0)
• Look for errors in: View → Output → GitLab AI Assistant

GitLab Connection Issues

• Verify GitLab URL is correct (include https://)
• Ensure token has api scope
• Test connection: GitLab AI: Configure → Test Configuration

LLM API Errors

• Check API key is valid
• Verify API URL for custom/Ollama providers
• Check model name is correct for your provider
• Review quota/rate limits

Code Review Not Working

• Ensure you have uncommitted changes
• Check LLM provider is responding
• Verify reviewMode setting is valid

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Run tests and linting
5. Submit a pull request

License

LICENSE

Support

For issues and feature requests, please file an issue on the repository.