What problem does it solve? Long Markdown documents produced by AI agents (plans, specs, system designs, postmortems) are hard to read as raw text walls, and nobody on the team wants to open a plain .md file. This Skill analyzes the document structure and renders it as a single self-contained HTML page that reads like a designed docs page. ## Core Features & Use Cases - Semantic component mapping: The agent decides which sections become Mermaid diagrams, step timelines, pros-cons tables, comparison cards, callouts, or collapsible panels instead of doing a literal Markdown-to-HTML conversion. - Self-contained output: One HTML file with embedded CSS, light/dark Claude-orange theme, sidebar TOC with scroll-spy, copy-to-clipboard code blocks, and KaTeX formula rendering via CDN. - Multi-language UI: Interface labels follow the source document language, with built-in label tables for English, Vietnamese, Chinese, Japanese, Korean, Spanish, French, and German. - Use Case: After your AI writes a 2,000-word migration plan, run the skill on plan.md to get plan.html with a Mermaid architecture diagram, a five-step timeline, risk callouts, and a comparison of message-bus options — ready to email or drop in Slack. ## Quick Start Ask your agent to run /md2html on a Markdown file such as plan.md, and it will write a styled plan.html next to the source and open it in your browser.
