Shift-V2: Real-Time Architectural Guidance Extension

Shift-V2 is a VS Code/Cursor extension that provides real-time, proactive guidance on architectural best practices. Unlike linters or reactive AI tools, Shift-V2 focuses on the big picture—analyzing your entire codebase to spot high-level design issues like N+1 queries or inefficient component usage. It delivers clear, actionable tips without you needing to ask.

Features

• Whole-Codebase Analysis: Monitors your project and identifies architectural inefficiencies using OpenAI's O3-Mini model and a graph database for precise context.
• Automatic Tips: Get alerts like "This might slow your database queries" as you save files.
• Tailored Fixes: Suggestions like "Fetch this data in one batch" that fit your code.
• Clear Explanations: Understand why it matters, e.g., "Multiple queries here can bottleneck your app."
• Ask Questions: Hit "Clarify" to dig deeper into any tip.

Example: A notification with a detected issue and option to see details.

Requirements

• VS Code: Version 1.96.0 or higher.
• OpenAI API Key: Required for the O3-Mini model. Get one from OpenAI.

Extension Settings

Shift-V2 adds the following setting:

• shift-v2.openaiApiKey: Your OpenAI API key for accessing the O3-Mini model. Set it in VS Code Settings.

How to Set Up

1. Install the Extension: Install Shift-V2 from the VS Code Marketplace or sideload the .vsix file.
2. Set API Key:
   • Open VS Code Settings (Ctrl+, or Cmd+, on Mac).
   • Search for shift-v2.openaiApiKey.
   • Enter your OpenAI API key.
3. Ensure Kùzu Integration: The extension uses a Kùzu graph database for enhanced analysis. No additional setup is required if installed via the official release.

How to Use

1. Start Coding: Open your project in VS Code. Shift-V2 automatically analyzes your codebase as you save files, leveraging the Kùzu database to examine code constructs (e.g., functions, classes) and their relationships (e.g., calls, data usage).
2. View Notifications: When an architectural issue is detected, a notification appears with a brief description (e.g., "N+1 query detected"). Click "See Details" for the full explanation and fix.
3. Act on Suggestions: Apply the suggested fix or click "Clarify" to ask the AI for more details.

Example

Suppose you write a function that fetches user data in a loop:

1: async function getUserPosts(users) {
2:   const posts = [];
3:   for (const user of users) {
4:     posts.push(await db.query("SELECT * FROM posts WHERE user_id = ?", user.id));
5:   }
6:   return posts;
7: }