kasuncfdo/ghost-headless-blog-skill

This project is a skill for Claude Code, meaning it is a packaged set of instructions and code templates that teach the AI assistant how to add a blog to a website. Specifically, it helps set up a headless Ghost CMS blog inside a Next.js website. Ghost is a popular blogging platform, and headless means the blog's content is managed in Ghost but displayed using your own website's design instead of Ghost's own theme system. Once installed, Claude automatically uses this skill whenever someone asks for something like adding a Ghost powered blog to a site, or when troubleshooting an existing Ghost blog that shows no posts, has stale content, or has broken images or cards. Instead of guessing how to wire this up, Claude follows tested patterns already written out in the skill. The skill includes reference documents covering setup steps like configuring Ghost's admin area and environment variables, page routing patterns, and how to look up Ghost's own documentation. It also ships ready to use code templates: a client for talking to Ghost's content API, image handling code for faster loading, a webhook handler that refreshes cached pages the moment new content is published, and matching visual styles for Ghost's content blocks. Some of the more advanced features baked in include full tag and author archive pages, author bios with social links, search engine friendly metadata, and a sitemap that lists posts, tags, and authors. It is also designed to fail gracefully: if the Ghost connection is misconfigured, the site still builds successfully and shows an empty blog rather than crashing. Installing the skill can be done a few ways, including as a Claude Code plugin, through a general skills installer command that works across several AI coding tools, or by manually copying the skill files into a project or user folder. Using the generated blog requires a Next.js 15 or newer project, an actual Ghost site with an API key, and hosting that supports content refreshing on demand, such as Vercel. The project is released under the MIT license, a permissive license allowing free use, modification, and redistribution, including commercially, as long as the copyright notice is kept.