Infineon ModusToolbox™ AI Assistant

Preview — Available in the VS Code Marketplace

The ModusToolbox™ AI Assistant brings hardware-aware intelligence to your AI coding experience. It gives GitHub Copilot (and other AI assistants like Cursor and Windsurf) deep understanding of your ModusToolbox™ workspace — your specific MCU, BSP, available libraries, and peripheral configurations — so you get relevant, compilable code instead of generic suggestions.

Why Use This Extension?

Without workspace context, AI assistants generate code that looks correct but targets the wrong device, uses deprecated APIs, or misses critical initialization steps. The ModusToolbox™ AI Assistant solves this by providing:

• Device-specific code generation — AI responses are grounded in your actual MCU, BSP, and pin configuration
• Library-aware suggestions — The assistant knows what middleware is installed and what's available to add
• Skill-guided workflows — Structured knowledge for peripherals (GPIO, I2C, SPI, UART, DMA), connectivity (WiFi, BLE, MQTT), RTOS patterns, and more
• Project creation assistance — Guided workflows to scaffold new ModusToolbox™ projects from templates and code examples
• Tool integration — Launch Device Configurator, Library Manager, and other ModusToolbox™ tools directly from the chat

Supported Devices

Device support is actively expanding. PSOC™ Edge coverage will continue to deepen alongside the extension as it moves toward general availability.

Getting Started

Install the ModusToolbox™ AI Assistant from the VS Code Marketplace. The ModusToolbox™ for VS Code extension is required and will be prompted for installation if not already present.

The extension supports two workflows, both accessible from the ModusToolbox™ for VS Code extension UI.

Flow A: AI-Assisted Project Creation

For new projects where you want AI to help plan and scaffold your application, use this flow. This flow is also useful for porting existing projects from one device to another - for example, from STM32 or PSOC™ 6 to PSOC™ Edge or PSOC™ Control.

1. In the ModusToolbox™ for VS Code extension, click Create Project
2. Choose "Use AI to Assist" (instead of "Guided Setup")
3. A temporary workspace opens with planning and project creation skills pre-loaded
4. Open Copilot Chat (it may open automatically) and describe what you want to build - provide a simple prompt, full requirements, or a link to an existing project (URL or directory)
5. The AI assistant drafts a project plan for your review
6. Approve the plan, and the agent automatically opens a new VS Code instance with the created project and begins executing the plan

Flow B: Add AI Enhancements to an Existing ModusToolbox™ Project

For ModusToolbox™ projects you've already created or opened:

1. Open your ModusToolbox™ project in VS Code
2. In the Application tab of the ModusToolbox™ for VS Code extension, click "Add AI Enhancements"
3. The extension deploys instruction files, skill files, and tool awareness into your workspace
4. Open Copilot Chat and start asking hardware-aware questions about your project

After Setup

Once either flow completes, you have full AI-assisted development. The assistant understands your hardware, libraries, and project structure.

Example prompts:

• "Create a FreeRTOS task that reads the BMI270 accelerometer over I2C and prints values over UART"
• "Set up BLE peripheral advertising with a custom GATT service"
• "What libraries are available for WiFi connectivity?"
• "Configure DMA for SPI transfers on this board"

How It Works

The extension gives AI assistants deep context about your ModusToolbox™ project through three layers:

Workspace Analysis

A local server analyzes your workspace and provides AI assistants with structured knowledge about your project:

The server runs entirely on your local machine. It fetches the Infineon super-manifest from GitHub on startup to resolve the board and library catalog; all other operations are fully offline.

AI Instruction Files

The extension generates instruction files that teach your AI assistant about the project:

All three files contain identical skill references and workspace context — they point to the same .github/skills/ directory deployed alongside them. This ensures the extension works across multiple AI-assisted development tools, including environments where MCP servers are restricted by IT policy.

Skill Files

The extension deploys structured skill files into your workspace that teach AI assistants proven patterns for:

• Peripheral initialization (GPIO, I2C, SPI, UART, ADC, PWM, DMA)
• Connectivity stacks (WiFi, BLE, MQTT, HTTP)
• FreeRTOS task management and synchronization
• Dual-core (CM33/CM55) project architecture
• Device Configurator workflows
• Power management and deep sleep

Skills provide hardware-specific recipes that produce correct, compilable code — validated across real hardware. As device family support expands, the extension will deploy only the skills applicable to your selected device.

Requirements

• VS Code 1.85 or later
• ModusToolbox™ 3.7 or later (latest release recommended)
• Python 3.9 or later
• ModusToolbox™ for VS Code extension (installed automatically as a dependency)

Troubleshooting

Disabling the AI Assistant

Disable this extension for your workspace from the Extensions view (Ctrl+Shift+X / Cmd+Shift+X). Find Infineon ModusToolbox™ AI Assistant and disable it for the workspace.

Optional cleanup — remove files created by the extension that may affect AI interactions:

• Delete .modustoolbox/docs/ directory
• Edit or remove extension-generated content from .github/copilot-instructions.md, .cursorrules, and .windsurfrules
• Delete .github/skills/mtb-* directories
• Delete files .github/instructions/mtb-*

Feedback & Support

This extension is in Preview. We welcome feedback to help shape the product:

• Report issues through your Infineon support channel
• Check the Output panel ("ModusToolbox AI Assistant") for diagnostic logs

License

Apache-2.0

Built by the Infineon ModusToolbox™ team.