What problem does it solve? Mermaid diagrams frequently fail to render because opening keywords, beta suffixes, and feature support vary across Mermaid versions pinned by GitHub, GitLab, VS Code, and other hosts. This Skill picks the correct diagram type, writes the required frontmatter and accessibility header, and validates the source before it ships. ## Core Features & Use Cases - Type selection by question: Maps the question a diagram must answer to the correct type and opening keyword across 30+ diagram types, with stability tiers (stable, beta-aliased, beta-only, external) for renderer compatibility. - Header and accessibility enforcement: Ensures every diagram carries a YAML frontmatter config: block plus accTitle/accDescr so screen readers and raw-Markdown readers can interpret it. - Static validation and rendering checks: The check-mermaid.mjs script scans Markdown files for unrecognised keywords, deprecated %%{init}%% directives, missing accessibility headers, and unbalanced quotes, and runs a true parse via mmdc when available. - Use Case: A README architecture diagram shows a parse error on GitHub but renders on mermaid.live. Use this Skill to identify that architecture-beta requires the beta spelling on older renderers, fix the header, and re-validate the file. ## Quick Start Ask the assistant to create a Mermaid sequence diagram of your login flow for a GitHub README and validate it before committing.