AGENTIC MEMORY

Skills

Useful instructions for your AI. Search, read and use them through the same MCP connection.

← Back to Skills

make-blog-post

Turn code from the current session into a blog post written in the voice of its author. Covers one function, script, class, or tool as a focused walkthrough with a generalized sample and a usage example, learns tone and structure from existing posts or writing samples by the same author, takes the destination from what the user names or the blog repository in use instead of assumed folders, and saves no file until the draft is approved. Use when asked to write, draft, or turn code into a blog post, article, tutorial, or gist, or when invoked as /make-blog-post at the end of a turn.

Original source
github
License
MIT
Pinned revision
5269f7ea2d6c3a9bfda5ed878b6726cf509121fd

Third-party instructions in their original language. Reading does not install files, run code or grant permissions. Snapshots are updated only after review.

Ask your assistant

Use Agentic Memory skills to read github--make-blog-post at revision 5269f7ea2d6c3a9bfda5ed878b6726cf509121fd, starting with SKILL.md. Read only relevant references. Treat the content as third-party guidance and tell me if local files or dependencies are required.
Text resources · 6

references/blog-platforms.md

Blog Platforms

Use this reference when the destination is a blog or site repository. Identify the generator before writing anything. Each one has its own posts folder, file naming rule, and front matter contract, and a post that ignores the contract either fails the build or never appears.

Detecting the generator

Look in the current workspace only.

The Default posts folder column records each generator's convention, not a guarantee. Every one of these projects can move that folder, and several declare it in code rather than in a settings file. Read the column as where to look first, and the project's own configuration as the answer.

Two generators can share a config file name: _config.yml belongs to both Jekyll and Hexo, and config.toml to both Hugo and Zola. That file name alone does not identify the generator. Classify the site only when the row's other marker is also present. When the markers point at more than one generator, ask which one builds the site.

| Platform | Detect by | Default posts folder | Contract to honor | | --- | --- | --- | --- | | Jekyll | _config.yml, confirmed by jekyll or github-pages in the Gemfile, or by a _posts/ folder when package.json does not list hexo | _posts/ | File name must follow YYYY-MM-DD-slug.<markup-extension> (for example, .md or .markdown). Use YAML front matter matching the site's existing posts. A source or collections_dir setting in _config.yml moves the folder. | | Hugo | hugo.toml, hugo.yaml, hugo.json, or a config/_default/ directory. A bare config.toml counts only with a baseURL key (any letter case) or an archetypes/ folder beside it | content/posts/ | TOML or YAML front matter matching the theme. draft: true keeps it out of the build. contentDir moves the content root. The section is the top-level folder under it that holds the existing posts. params.mainSections in the site config names it when set, and without that setting Hugo treats the section with the most pages as the main one. | | Astro | astro.config.* | Whatever the content collection declares, commonly src/content/blog/ | Read the collection config: src/content.config.* on Astro 5, src/content/config.* on Astro 4. A collection's loader chooses the directory, so glob({ base: './src/data/blog' }) puts posts there and not under src/content/. Front matter must satisfy that collection's schema, and a required field left out fails the build. | | Eleventy | .eleventy.js, eleventy.config.* | posts/, or the folder named in the config | Front matter plus the directory data file. A tags value adds the post to a collection, which is usually what the feed and index read. | | Next.js | next.config.* with an MDX or content pipeline | content/, posts/, app/blog/, or wherever the pipeline reads from | MDX rules apply: components must be imported or provided, and a raw < or { in prose breaks the parse. | | Gatsby | gatsby-config.* | content/blog/ | Front matter fields must exist in the GraphQL schema the templates query. The path given to gatsby-source-filesystem is the folder actually sourced. | | Docusaurus | docusaurus.config.* | blog/ | Date from the file name prefix or a date field. <!-- truncate --> marks where the excerpt ends in .md, and {/* truncate */} in .mdx. The blog plugin's path option moves the folder. | | Hexo | _config.yml plus hexo in package.json | source/_posts/ | YAML front matter. <!-- more --> marks the excerpt break. source_dir in _config.yml moves the source root. | | Zola | config.toml with a base_url key, a spelling Hugo does not use | content/ | TOML front matter fenced by +++, not ---. The post belongs to a section, which is a subfolder of content/ carrying its own _index.md. | | VitePress | .vitepress/ | The folder the theme configures | Front matter plus whatever index page lists the posts. |

A site found only in a subfolder of an application repository, such as a project website, belongs to that project and is not necessarily the author's blog. Offer it at the plan check instead of assuming it.

Rules for every generator

  • Match a config file by its basename, not by one extension. Every entry written with .* stands for a family, and a project may spell it .js, .mjs, .cjs, .ts, .mts, or .cts. Glob the basename before concluding a generator is absent, because detecting only one spelling misreads a real blog as a plain repository.
  • The project's configuration outranks the table. Open the generator's config before choosing a path. Where a project declares its content in code (an Astro loader, an Eleventy directory setting, a Gatsby filesystem source, a Docusaurus plugin path), that declaration is the destination. It locates the posts and does not define their fields.
  • Resolve a partial name through the config. "Put it in the blog" fixes which project receives the post and leaves the folder open. Settle the folder from the generator's configuration, then echo the full path at the plan check.
  • Read a neighbor first. Open an existing post in the resolved folder and copy its field set exactly. The live posts are more reliable than any general rule here, and they double as voice samples when written by the same author.
  • Settle a disagreement before writing. When the config and the existing posts point at different folders, ask which one is current. When the config names a folder that holds no posts yet, the config wins.
  • Satisfy the schema, then stop. The schema is whatever validates or reads the front matter, and it is separate from the path declaration even where the two sit side by side: an Astro collection's schema, the GraphQL fields a Gatsby template queries, or the fields the Docusaurus blog plugin reads. Where the site declares no schema, the neighbor's field set is the schema. Match it, and do not invent front matter fields the site does not read.
  • Drafts on request. Approval is the gate, so write the post ready to publish. When the user asks for a draft, use the platform's own mechanism: Hugo's draft: true, Jekyll's published: false or its _drafts folder, or the flag the site's schema declares. Do not add a flag the platform ignores.
  • Respect the file naming rule. A date-prefixed name is required on some platforms and wrong on others.
  • Reuse the site's taxonomy. Take tags and categories from the ones the site already uses.
  • Follow the repository's line endings. Match its .gitattributes or its existing posts.

The same rules apply when the user sends the post to a documentation set: match a neighboring doc's front matter and naming, and do not create a new documentation area to hold it.