This project provides an open-source collection of "skills" enabling AI coding assistants to directly control embedded development tools like compilers, debuggers, and communication buses. It targets engineers and power users, automating the full workflow from code generation to hardware verification, creating a closed-loop AI development cycle.

How It Works

embeddedskills encapsulates command-line tools (e.g., Keil, GCC, J-Link, OpenOCD, Wireshark) into Python-based "skills." Each skill exposes its functionality via SKILL.md files, allowing AI assistants supporting the Skill protocol to invoke them like functions without retraining. Skills return structured JSON output detailing status, results, and next steps, empowering AI to autonomously manage compilation, flashing, debugging, and error correction.

Quick Start & Requirements

Installation is straightforward via npx: npx skills add https://github.com/zhinkgit/embeddedskills -g -y. Key requirements include an AI assistant compatible with the Skill protocol and specific toolchains/debuggers depending on the skill used (e.g., Keil MDK, CMake, ARM GNU Toolchain, SEGGER J-Link, OpenOCD, probe-rs, Wireshark). Most skills rely on Python's standard library.

Highlighted Details

• Supports diverse AI assistants (Claude Code, Copilot, TRAE, Cursor, Kiro, etc.) via the Skill protocol.
• Automates the complete embedded lifecycle: code generation, compilation, flashing, and hardware debugging.
• Enables orthogonal combinations of build systems (Keil/GCC) and debuggers (J-Link/OpenOCD/probe-rs).
• Integrates seamlessly with existing projects without requiring modifications.

Maintenance & Community

The project is actively maintained, open-source under the MIT license, and welcomes community contributions via Issues and Pull Requests. It benefits from support from the Linux.do community.

Licensing & Compatibility

Licensed under the permissive MIT license, it generally allows for commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

PlatformIO and pyocd skills are pending support. The CAN skill requires further testing. On Windows, using probe-rs with J-Link may necessitate switching the J-Link driver to WinUSB, potentially affecting SEGGER's official tools.