feat: add blog content and update preview logic

[americano98]

Summary

This PR moves published blog markdown into the website repository and makes content/blog the canonical source for production blog content. Normal blog pages now read markdown from the local filesystem instead of the external blog-content repo or CDN. Media assets remain on the CDN and are referenced by URL from markdown frontmatter/content.

Full preview index/category snapshots load from the actual GitHub branch tree at the resolved commit. The loader reads the recursive Git tree, selects only content/blog authors/categories/posts, and fetches file text in parallel GraphQL blob batches. This avoids per-post REST Contents API fan-out, avoids GitHub compare file-list caps, and keeps preview content faithful to the selected branch even when it is behind main. Preview routes also set a longer function duration for cold full-branch loads. Single-post preview remains scoped to the requested markdown file.

The PR also adds a Sync Main GitHub Actions workflow intended for the private mirror repository. The workflow can keep the mirror main aligned with the public upstream main while leaving author branches untouched.

Notes

• content/blog/posts, content/blog/authors, and content/blog/categories are now committed to this repo.
• The old CDN/blog-content markdown source is removed from the runtime path.
• Preview mode does not use a preview secret; access is controlled by environment configuration and the private repo token.
• The mirror sync workflow assumes the upstream repository is publicly readable.

Validation

• blog post preview

[vercel]

@americano98 is attempting to deploy a commit to the neondatabase Team on Vercel.

A member of the Team first needs to authorize it.