PostKing Blog

Handles the blog lifecycle on PostKing: publications (the containers articles live under), AI article generation, review/editing, and publishing — either to the PostKing-hosted blog or out to connected external platforms. Requires an active brand — use the postking skill's brand-pick flow first if one isn't set.

When to Use

Use this skill when the user wants to create or manage a blog publication; generate, review, edit, or rewrite a blog article; publish an article (hosted or external); check on an in-progress generation; or manage blog authors/categories.

When NOT to Use

• Social posts, content weeks, or repurposing to social → postking-social.
• Landing pages or side pages → postking-landing-pages.
• The SEO/GEO keyword-to-roadmap pipeline (seeds → keywords → clusters → briefs) that eventually produces briefs for articles → postking-seo. This skill covers direct/manual blog generation and publishing, not that pipeline's brief-to-article automation.
• Saved voice profile management or a standalone de-slop pass → postking-brand-voice.
• No active brand yet → postking-getting-started.

Minimal tool subset

• list_publications, create_publication, update_publication, delete_publication, list_blogs — publications.
• generate_blog_post, get_blog_status, get_job — async generation + polling.
• get_blog_article, update_blog_article, rewrite_text — review and iterate.
• list_assets, import_asset_from_url — pick or import an image to set as an article's featured/header image.
• publish_blog_article, list_publishing_connections, schedule_blog_article — publishing.
• list_blog_authors, create_blog_author, list_blog_categories, create_blog_category — metadata.
• import_blog_articles, delete_blog_article — housekeeping.

Procedure

Set up a publication

1. list_publications({ brandId? }) (or list_blogs, which also returns a statusBreakdown of existing articles) — find an existing publication.
2. create_publication({ title, description?, layout?, languageCode?, brandId? }) if none exists — returns a publicationId. languageCode is optional (defaults to the brand's content language) and must already be enabled on the brand's language roster — see the postking skill's language section; an un-enabled code is rejected with a 403.
3. delete_publication({ publicationId, brandId? }) to remove one — only allowed when it has zero articles and no connected domain/external sync/publishing connections; the API refuses with an explanation of what's still attached otherwise (delete every article first with delete_blog_article, disconnect domain/connections).

Generate a blog article

1. generate_blog_post({ publicationId, topic, voiceProfileId?, targetLength?, primaryKeywords?, secondaryKeywords?, brandId?, language? }) — note the param names: primaryKeywords/secondaryKeywords, not keywords; targetLength (short/medium/long), not length. language (en/es/pt-BR/de/fr/cs) overrides the brand's contentLanguage for this article only. Async; returns articleId + operationId.
2. Poll get_blog_status({ articleId }), or get_job({ pollUrl: operationId, brandId, wait: true }), until status is completed (or failed).
3. get_blog_article({ articleId }) — review the draft (full content, SEO fields, previewUrl, editUrl, and the current featuredImageUrl/featuredImageAlt/featuredImageDescription).

Review and iterate

• update_blog_article({ articleId, title?, content?, excerpt?, status?, metaTitle?, metaDescription?, authorId?, categoryId?, featuredImageUrl?, featuredImageAlt?, featuredImageDescription?, cta?, sidePageInfo? }) for direct edits — only the fields you pass are changed.
• rewrite_text({ text, voice?, platform: "blog" }) to polish a specific passage, then paste the result back in via update_blog_article.
• Re-run generate_blog_post with a refined topic/primaryKeywords for a fresh pass instead of patching piecemeal changes.

Set or replace the featured (header) image

The header image can be set at generation time (generate_blog_post's generateAiImage/attachVisualAsset/selectedAssetId), but to add or swap it on an existing article:

1. Get an image URL — list_assets({ brandId? }) and copy an asset's url/fileUrl (returned as an absolute CDN URL, e.g. https://cdn.postking.app/assets/<brandId>/images/x.png); or import_asset_from_url({ url }) to pull an external image into the brand library first; or use any public image URL / data: URI directly.
2. update_blog_article({ articleId, featuredImageUrl, featuredImageAlt?, featuredImageDescription? }) — sets the header image. Pass featuredImageUrl: "" to remove it. An external URL is auto-downloaded to brand storage when the article is published.
3. get_blog_article({ articleId }) — the response echoes featuredImageUrl/featuredImageAlt/featuredImageDescription, so you can confirm the change stuck.

Edit the CTA (call-to-action) block

The CTA lives in structured data (sidePageInfo on the article), not in the article body — never write CTA markup into content to change it; that gets stripped/ignored.

• Simple case — update_blog_article({ articleId, cta: { url, label?, headline?, body? } }):
  • url → CTA link target (required when enabling a CTA)
  • label → CTA button text
  • headline → CTA block headline
  • body → CTA block body copy
• Disable/remove the CTA — update_blog_article({ articleId, cta: { enabled: false } }) (equivalent to sidePageInfo: null).
• Advanced case — link the CTA to an existing side page instead of a raw URL: update_blog_article({ articleId, sidePageInfo: { id } }) (id from list_side_pages). cta and sidePageInfo are mutually exclusive — pass only one per call.
• Example:

  update_blog_article({
    articleId: "art_123",
    cta: {
      url: "https://example.com/pricing",
      label: "See pricing",
      headline: "Ready to get started?",
      body: "Compare plans and find the right fit."
    }
  })
  

• Malformed CTA payloads are now rejected server-side (strict validation) instead of being silently persisted — pass exactly the documented fields (misspelled keys, e.g. ctaUrl instead of url, will error rather than silently doing nothing).

Publish

• PostKing-hosted: update_blog_article({ articleId, status: "published" }) makes it live on the brand's PostKing blog immediately.
• Scheduled hosted publish: schedule_blog_article({ articleId, scheduledAt }) — future ISO 8601 datetime; auto-publishes (and auto-pushes to any autoPublish-flagged external connections) when the time arrives.
• External (WordPress / Medium / Substack, connected via the dashboard): list_publishing_connections({ publicationId }) to find connection IDs, then publish_blog_article({ articleId, connectionIds: [...] }).
• get_blog_status({ articleId }) — confirm the final status and get the live URL(s).

CLI fast path

For the full command catalog, use the postking skill's references/commands.md, or run pking blogs --help / pking publications --help.

Pitfalls

• generate_blog_post param names: primaryKeywords/secondaryKeywords and targetLength, not keywords/length. Passing the wrong names silently drops them (extra zod fields are ignored, not errored).
• Hosted publish vs. external publish are different tools. publish_blog_article requires a non-empty connectionIds array — it's for pushing to external platforms, not for making the PostKing-hosted copy live. Use update_blog_article({ status: "published" }) for that.
• No pking blogs auto-assign-cta CLI command exists. If the user wants a CTA auto-assigned to a piece of content, that capability is the MCP tool seo_auto_assign_cta (owned by the postking-seo skill's surface) — don't invent a CLI equivalent.
• Generation is async and can take a couple of minutes. Poll get_blog_status or get_job rather than assuming failure early; get_job supports wait: true to block instead of polling manually.
• update_blog_article is a partial update. Omitted fields are left unchanged — no need to resend the whole article to change one field.
• The CTA is structured data, not article HTML. Use cta/sidePageInfo on update_blog_article, never embed CTA markup in content — it will be stripped on write, not rendered. Malformed cta/sidePageInfo shapes (unknown keys, empty {}, enabling without a url) are rejected with a 400, not silently dropped.
• The featured image is a URL/path, not an asset ID. featuredImageUrl takes an image URL, a brand-asset URL (the fileUrl from list_assets — now an absolute CDN URL, e.g. https://cdn.postking.app/assets/<brandId>/images/x.png), or a data: URI — not an asset ID. Bare legacy paths (/assets/<brandId>/...) are still accepted as input for backward compatibility. (Contrast generation, where generate_blog_post uses selectedAssetId, an asset ID.) Pass an empty string to clear it.

Verification

• get_credits({ detail: "short" }) — confirms auth and balance.
• list_publications({}) — returns at least one publication once set up (or an empty list if none created yet, which is expected pre-setup).