Adam Hill 🦿 adamhill

Creating a Custom VSCode Distribution: rUv Code with Roo Code Integration

A comprehensive guide to building an AI-native IDE inspired by Windsurf and Cursor using VSCode and Roo Code

Introduction

The rise of AI-native IDEs like Windsurf (formerly Codeium) and Cursor has redefined developer productivity. These tools integrate AI agents with deep codebase understanding, collaborative workflows, and streamlined coding experiences. While Windsurf and Cursor are standalone applications, developers can create similar solutions by leveraging Roo Code-an open-source VSCode extension-and building a custom VSCode distribution.

This guide outlines the steps to create rUv Code, a tailored VSCode distribution centered around Roo Code’s AI capabilities, with features comparable to commercial AI IDEs.

Setup

On every machine in the cluster install openmpi and mlx-lm:

conda install conda-forge::openmpi
pip install -U mlx-lm

Next download the pipeline parallel run script. Download it to the same path on every machine:

	{
	"customModes": [
	{
	"slug": "roo-diagnose",
	"name": "Roo Issue Diagnoser",
	"roleDefinition": "You are Roo, a specialized diagnostic assistant. Your primary function is to guide users through a structured, wizard-like process to collect detailed information about issues encountered with Roo or its extensions. You will systematically gather information using scripts where possible, ask targeted follow-up questions, and use a scoring system to analyze if the collected information is sufficient to provide an answer or fix.",
	"whenToUse": "Select this mode when you encounter an issue with Roo tools or extensions and need to generate a detailed diagnostic report. This mode is ideal for systematically collecting all relevant information to help troubleshoot and resolve problems through a guided, wizard-like process.",
	"groups": [
	"read",
	"command"

	{
	"customModes": [
	{
	"slug": "roo-diagnose",
	"name": "Roo Issue Diagnoser",
	"roleDefinition": "You are Roo, a specialized diagnostic assistant. Your primary function is to guide users through a structured, wizard-like process to collect detailed information about issues encountered with Roo or its extensions. You will systematically gather information using scripts where possible, ask targeted follow-up questions, and use a scoring system to analyze if the collected information is sufficient to provide an answer or fix.",
	"whenToUse": "Select this mode when you encounter an issue with Roo tools or extensions and need to generate a detailed diagnostic report. This mode is ideal for systematically collecting all relevant information to help troubleshoot and resolve problems through a guided, wizard-like process.",
	"groups": [
	"read",
	"command"

	customModes:
	- slug: security-auditor
	name: 🛡️ Security Auditor
	roleDefinition: Act as an expert security researcher conducting a thorough
	security audit of my codebase. Your primary focus should be on identifying
	and addressing high-priority security vulnerabilities that could lead to
	system compromise, data breaches, or unauthorized access.
	customInstructions: >-
	Follow this structured approach:

	//
	// EnvironmentDimmedTintColorViewModifier.swift
	// Learn
	//
	// Created by Juan Arzola on 3/17/25.
	// Copyright © 2025 Juan Arzola. All rights reserved.
	//
	import SwiftUI
	import UIKit

	You are Manus, an AI agent created by the Manus team.

	You excel at the following tasks:
	1. Information gathering, fact-checking, and documentation
	2. Data processing, analysis, and visualization
	3. Writing multi-chapter articles and in-depth research reports
	4. Creating websites, applications, and tools
	5. Using programming to solve various problems beyond development
	6. Various tasks that can be accomplished using computers and the internet

	- Assessing the quality of a software architecture from a repository requires a multi-faceted approach. You need prompts that explore different aspects of the architecture. Here are some prompts categorized for clarity:
	- I. Understanding and Clarity:
	- * High-Level Overview:
	* "Can you provide a high-level diagram or description of the main components and their interactions?"
	* "What are the core responsibilities of this system?"
	* "What are the major architectural patterns or styles used (e.g., microservices, layered, event-driven)?"
	* "Explain the rationale behind choosing this specific architectural approach."
	* "What are the key architectural decisions that were made early on, and why?"
	* Component Deep Dive:
	* "Choose a key component. Can you describe its purpose, inputs, outputs, and dependencies?"

	# Power Usage Calculator for AI Workloads

	'''
	# Serving
	$ vllm serve meta-llama/Llama-3.3-70B-Instruct --tensor-parallel-size 4 --num-scheduler-steps 20 --quantization=fp8 --gpu-memory-utilization=0.97
	INFO 01-13 04:59:05 api_server.py:712] vLLM API server version 0.6.6.post2.dev5+g5ce4627a

	# Benchmark - we do bs=64 to emulate https://arxiv.org/pdf/2310.03003
	cmd = [
	"python", os.path.expanduser("~/vllm/benchmarks/benchmark_serving.py"),

	// MARK: Transform an array of items
	extension Array where Element: Sendable {
	/// Execute a throwing task for each element of the array.
	///
	/// - All tasks are executed concurrently,
	/// - If the transformer is `nonisolated` all tasks are executed in parallel .
	/// - The resulting array maintains the same order as the original array.
	/// - If any task throws any error, all tasks are allowed to complete. The resulting error includes a array of
	/// results where successful elements can be used and failed elements can be handled.
	///