agentleFS
kevinbadikevinbadi/seo-agent-kit/.claude/skills/wordpress-blog/SKILL.md

wordpress-blog

Skill9 starsChanged 42 days agoMITView on GitHub

Connect a WordPress site (self-hosted or WordPress.com) through Creator OS and list, write, update, publish or delete blog articles on your own domain from Claude (MCP tools, REST API or the creatoros CLI). Use when the user wants to connect their blog, write or edit one specific article, repurpose a winning social post into an SEO article, refresh an old post, bulk-fix every post, or publish an approved draft. For the automated daily loop use the seo-engine skill.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 3 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the API key of the workspace that connected the blog (a key is pinned to one workspace).

    line 31
  2. Very specific:

    Check it worked: GET /v1/blog / blog_list_sites.

    line 40
  3. Very specific:

    Write 1,100 to 1,700 words of HTML: short intro, a worked example (prompt, command, workflow), 4+ <h2> sections, an <h2>Frequently asked …

    line 68
  4. Very specific:

    Only state product facts you can point to (docs, pricing page, the seo-engine facts.json).

    line 71
  5. Distinctive:

    No invented stats, customers or quotes.

    line 72
  6. Very specific:

    Use this kit's tables: which blog posts people entered on and then signed up from (seo_metrics_daily once the seo-engine measure job runs, …

    line 82
  7. Distinctive:

    Write 1,100 to 1,700 words of HTML following "Write one article now".

    line 108

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

WordPress blog via Creator OS

Creator OS writes to your WordPress blog for you. The articles live on your domain (good for SEO); your app does not have to run on WordPress. Three ways to drive it, same backend:

WayUse it when
MCP tools (blog_list_sites, blog_list_articles, blog_get_article, blog_create_article, blog_update_article, blog_delete_article, blog_get_connect_link)Claude / Claude Code / ChatGPT with the Creator OS connector (https://mcp.creatoros.ca/mcp)
REST: https://creatoros-production-5658.up.railway.app/v1/... with Authorization: Bearer $CREATOROS_API_KEYscripts, cron jobs, the seo-engine
CLI: npx @creatoros/cli blog:...terminal

Get the key at https://www.creatoros.ca (Settings > API keys). A key is pinned to one workspace, so use the key of the workspace that connected the blog.

Critical

  • Drafts by default (isPublished: false) unless the user said to publish. Ask before publishing or deleting anything when a human is in the loop.
  • Only state product facts you can point to (docs, pricing page, the seo-engine facts.json). No invented stats, customers or quotes. No em dashes. Describe competitors neutrally.
  • One distinct search intent per post. Publishing 10+ near-identical posts a day gets a site demoted.
  • Articles are HTML (bodyHtml), not Markdown.
  • Use the API key of the workspace that connected the blog (a key is pinned to one workspace).

1. Connect the site (once)

  • Self-hosted: WordPress admin > Users > Profile > Application Passwords > create one. Then POST /v1/connect/wordpress with { "site_url": "https://yoursite.com", "username": "...", "application_password": "abcd efgh ..." } or creatoros blog:connect-selfhosted --site https://yoursite.com --username you --password "abcd efgh ...".
  • WordPress.com: GET /v1/connect/wordpress (or MCP blog_get_connect_link) returns an auth_url; send the user there.
  • Check it worked: GET /v1/blog / blog_list_sites.

Tip: run WordPress on a subpath (yoursite.com/blog) behind your main app, or on a subdomain, and render the posts in your own design through the WordPress REST API (/blog/wp-json/wp/v2/posts). The ai-search-files skill builds llms.txt, sitemaps and FAQ schema from those same posts.

2. Endpoints

GET    /v1/blog/articles?limit=50&cursor=...   list (published + drafts): id, handle, title, status, tags
GET    /v1/blog/articles/{id}                  one article with bodyHtml
POST   /v1/blog/articles                       { title, bodyHtml, handle, excerpt, tags[], image{url,altText}, isPublished }
PATCH  /v1/blog/articles/{id}                  any of the above, e.g. { "isPublished": true }
DELETE /v1/blog/articles/{id}
  • Articles are HTML (bodyHtml), not Markdown.
  • handle is the slug. excerpt is what most themes use as the meta description (150-160 chars).
  • Yoast / Rank Math fields are not set through the API.
  • Drafts by default (isPublished: false) unless the user said to publish. Ask before publishing or deleting anything when a human is in the loop.

3. Workflows

Write one article now

  1. blog_list_articles so the new post covers new ground and can link to old ones.
  2. Pick ONE search intent. The keyword goes in the title (near the start, under 70 chars), the first paragraph, one <h2> and the handle.
  3. Write 1,100 to 1,700 words of HTML: short intro, a worked example (prompt, command, workflow), 4+ <h2> sections, an <h2>Frequently asked questions</h2> with 3-5 <h3> questions, and a Get started section linking the signup page. Link 3-6 existing posts.
  4. Only state product facts you can point to (docs, pricing page, the seo-engine facts.json). No invented stats, customers or quotes. No em dashes. Describe competitors neutrally.
  5. blog_create_article as a draft, give the user the link, publish on approval with blog_update_article { isPublished: true }.

Turn a winning social post into an article "Look at my three best-performing posts this month and turn the strongest one into a 1,200-word SEO article for my blog. Save it as a draft." Pull post analytics (Creator OS get_post_analytics or the PostHog sources card), expand the post's one idea into the article above, link the post.

Find what to write next (from PostHog) Use this kit's tables: which blog posts people entered on and then signed up from (seo_metrics_daily once the seo-engine measure job runs, or the Traffic sources card). Write distinct-intent siblings of the winners: a comparison, a niche version, a how-to, a model pairing.

Refresh an old post blog_get_article, update facts, add an FAQ section if missing, keep the handle, blog_update_article. Fresh modified dates get re-crawled; ping IndexNow after (see ai-search-files).

Bulk fix (every post) Loop GET /v1/blog/articles and PATCH each. Dry-run first and print what would change. Example: .claude/skills/seo-engine/scripts/add-youtube-links.mjs adds your YouTube link to every post.

Examples

Example 1: Connect a self-hosted blog User says: "Connect my WordPress blog at example.com" Actions:

  1. Have the user create an Application Password in WordPress admin.
  2. POST /v1/connect/wordpress (or creatoros blog:connect-selfhosted) with site URL, username and password.
  3. Verify with blog_list_sites. Result: The site is listed and ready for articles.

Example 2: Write one article as a draft User says: "Write a blog post about scheduling Instagram posts with AI" Actions:

  1. blog_list_articles to avoid overlap and find posts to link.
  2. Write 1,100 to 1,700 words of HTML following "Write one article now".
  3. blog_create_article with isPublished: false, share the link, publish on approval. Result: Draft on the user's blog, published only after approval.

Example 3: Refresh an old post User says: "Update my 2024 pricing post with current facts" Actions:

  1. blog_get_article, update facts, add an FAQ if missing, keep the handle.
  2. blog_update_article, then ping IndexNow (see ai-search-files). Result: Refreshed post with a new modified date.

Troubleshooting

Symptom: blog_list_sites / GET /v1/blog shows no site Cause: The WordPress site is not connected to this workspace, or the key belongs to another workspace. Solution: Connect the site (section 1) with the key of the workspace that owns it.

Symptom: only the first page of articles comes back Cause: The list endpoint pages with nextCursor. Solution: Loop until nextCursor is empty.

Symptom: SEO title / meta fields not set Cause: Yoast / Rank Math fields are not set through the API. Solution: Put the meta description in excerpt (150-160 chars), which most themes use.

Symptom: ranking drops after a burst of posts Cause: Near-identical posts or links to pages that don't exist. Solution: One distinct intent per post; only link pages that exist (your sitemap + the article list).

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.