Import Markdown File

Convert any raw markdown file into a sitemd page — generating frontmatter, reformatting content, determining group membership, and writing the page file.

Input

The user provides a file path as the argument:

$ARGUMENTS

If no argument is provided, ask the user for the path to the markdown file they want to import.

Procedure

1. Read the source file. Read the raw markdown content at the provided path. If it has existing frontmatter, extract any useful fields (title, description) but the frontmatter itself will be replaced.

2. Gather site context. Call these MCP tools in parallel:

   • Glob

     pages/**/*.md

     — get existing pages to avoid slug collisions and understand structure
   • Read

     settings/meta.md

     — get the site name and brand
   • Read

     settings/groups.md

     — get available groups for categorization

3. Determine metadata:

   • Title — Extract from the first

     #

     heading in the content, or derive from the filename. Do not append a site title suffix — the build engine adds it automatically.
   • Description — Write a concise summary of the page content, under 160 characters. SEO-friendly.
   • Slug — Derive from the title, kebab-cased. If assigning to a group, prefix with the group name (e.g.

     /docs/my-page

     ). Check against existing pages for uniqueness.
   • Group membership — Based on the content's topic, decide if it fits an existing group. Documentation-style content →

     docs

     . If no group fits, leave it as a standalone page.

4. Convert content:

   • Strip any existing frontmatter from the source
   • Convert YouTube/Vimeo

     <iframe>

     embeds to

     embed: URL

     syntax
   • Convert raw HTML image tags to markdown

     ![alt](src)

     where practical
   • Preserve all standard markdown as-is (bold, italic, code blocks, tables, lists, blockquotes, links, images)
   • Do NOT add the auto-injected syntax comment block — the build system handles that

5. Create the page. Call

   sitemd_pages_create

   with:

   • title

     ,

     description

     ,

     slug

     from step 3

   • content

     — the converted markdown body

   • groupMember

     — array of group names if assigned (e.g.

     ["docs"]

     )

   • sidebarGroupShown

     — same as group name if assigned

6. Report results:

   • File path created
   • Slug / URL
   • Group assignment (if any)
   • Any warnings (unresolved local image paths, content that couldn't be auto-converted)
   • Remind the user that the dev server auto-rebuilds, or suggest

     /launch

     if it's not running

Rules

• No permission required — Execute immediately without asking
• Never overwrite — If the target file already exists, report the conflict and ask the user what to do
• Preserve content — The goal is faithful conversion, not rewriting. Keep the author's voice and structure.
• Frontmatter field order — title, description, slug, sidebarGroupShown, groupMember
• Match conventions — Look at existing pages' title format, description style, and slug patterns