CAN bus reverse engineering with AI (Claude Code)

[MatinF]

showcase-combined-can-bus-reverse-engineering-ai.mp4

Hi all,

Recently I've published a new intro article on reverse engineering CAN bus data with AI - using a Claude Code skill. I'd love to hear your thoughts/feedback!

This is a direct sequel to my original intro to CAN bus reverse engineering, focused on the basic methodology i.e. the 'human approach', using e.g. webCAN and SavvyCAN .

With the release of our new CANsub CAN bus interfaces, I wanted to do a modern stab at this by developing a Claude Code skill around the CANsub and python-can.

I think the result is extremely effective - even if the skill is just a rough version. If you've used SavvyCAN in the past for this type of workflow I definitely recommend trying this out.

Some quick comments:

1. The skill is 100% open source, you can use/modify it as you see fit - and you're welcome to contribute to the skill on github with PRs

2. In the article you'll find a link to some of the sample data in case you want to try it out right away (you can also directly download it here if you prefer skipping the newsletter signup).

The data pack includes the data behind my vision OCR showcase, in case some of you e.g. want to attempt to use it to reverse engineer the turn signals or something similar.

3. If you find a cool use case for the skill and do e.g. a video about it, share it with me as I'd love to include some community showcases in the intro as well!

I hope you find this interesting and useful!

Martin, co-owner at CSS Electronics