LLM Optimized Output

[mrns]

In a local fork I have modified the skill so that it first produces an MD document and then the HTML visual explanation based on it. The MD document is the source of truth, meaning the HTML is based on the content generated on the MD document. These 2 artifacts serve different purposes though, the HTML is optimized for human readers while the MD is best suited for AIs consuming these docs. I also added a subcommand called update (eg: /visual-explainer update some-feature) which looks up an existing MD on the output directory under the given name (eg: some-feature.md), review latest version of the code and update both the MD and HTML accordingly.

Is this something you find interesting? Let me know and I can push a PR.