Claude Blog — a full content team inside Claude Code
A blog brings free visitors who become customers — but posts that actually rank take research, structure, SEO and editing. Claude Blog turns Claude Code into a content team that plans, writes, scores and optimizes posts that rank on Google and get cited by AI. This academy teaches you from scratch — no writing or tech skills required.
What is Claude Blog?
One command researches, outlines, drafts, fact-checks, adds schema and internal links — then scores the result and iterates until it's good.
Research first
Pulls SERP data and recent discourse so posts answer what people actually search and say.
Write & score
Drafts a full post, then runs a 5-gate quality contract, iterating until it scores 90+ with zero critical issues.
SEO & schema built-in
Titles, meta, headings, internal links and JSON-LD schema are handled automatically.
AI citations (GEO)
Optimizes posts to be quoted by ChatGPT, Gemini and Google's AI answers — not just ranked on Google.
Multilingual
Write once, then translate + localize + add hreflang across languages in one command.
Topic clusters
Plan a pillar page plus supporting posts so your whole blog builds authority.
🧠 How it works
Claude Blog is a full-lifecycle content engine. The orchestrator routes your request to the right sub-skill; for /blog write it runs a research-to-publish pipeline, scoring the draft against a 5-category 100-point system and iterating until it passes — backed by 5 specialist agents.
The audit pipeline
- 1 · Load brand & voiceReads BRAND.md / VOICE.md so the draft sounds like you, not generic AI.
- 2 · Researchblog-researcher finds current (2025–26) statistics, verifies sources, and pulls real discourse from Reddit/X/YouTube/HN.
- 3 · OutlineSERP-informed H2/H3 hierarchy with per-section word counts and content-gap analysis.
- 4 · Draftblog-writer writes answer-first content with sourced stats, citation capsules and a Key-Takeaways box.
- 5 · SEO + schemablog-seo validates on-page SEO; JSON-LD schema (BlogPosting, FAQ, etc.) is generated.
- 6 · Fact-checkEvery statistic is re-fetched from its cited URL to confirm it actually says that.
- 7 · Score & iterateblog-reviewer runs the 5-gate 100-point contract; rewrites until score ≥ 90 with zero criticals (max 3 passes).
- 8 · DeliverOutputs 8 publish-ready artifacts in one folder, dropping straight into your CMS.
skills/blog/SKILL.md routes every /blog request to the right sub-skill and runs the write pipeline.
One SKILL.md per job — write, rewrite, analyze, cluster, multilingual, schema, media, and more.
researcher, writer, seo, reviewer, translator — the workers behind the pipeline.
A 100-point score across content, SEO, E-E-A-T, technical and AI-citation readiness — the bar every post must clear.
🗺️ Your learning path
Work through the stages in order. Each lesson explains the idea in plain language first, then shows the exact commands. Mark lessons done to earn XP, level up, and unlock your certificate.
Orientation
Why a blog matters and what Claude Blog does for you.
⚙️ Install it on your machine
Claude Blog runs inside the Claude Code CLI. Pick the method that matches your setup — the plugin route is easiest.
/plugin marketplace add AgriciDaniel/claude-blog
/plugin install claude-blog@agricidaniel-claude-blogRun these inside a Claude Code session (1.0.33+).
/blog google data. Works with Next.js MDX, Astro, Hugo, Jekyll, WordPress, Ghost, 11ty, Gatsby and plain HTML.📦 Skills library
Every one of the 30 sub-skills, explained — what it does, when it kicks in, and a real example. Filter by area or search.
Full-lifecycle blog engine with 30 sub-skills, 12 templates and a 5-category 100-point score. Routes your request to the right place.
Triggers on: Any content request.
Write a new article from scratch optimized for Google + AI citations: template selection, answer-first format, Key-Takeaways box, citation capsules; iterates to 90+.
Triggers on: write a post · new article · blog about
SERP-informed outline: H2/H3 hierarchy, per-section word counts, content-gap analysis, FAQ planning, internal-linking zones.
Triggers on: outline · skeleton · structure a post
Full rewrite of an existing post for both Google rankings (E-E-A-T, latest core update) AND AI citations. Must score ≥ the original.
Triggers on: rewrite · optimize post · update content
Score one post on the 5-category 100-point system (content, SEO, E-E-A-T, technical, AI-citation) + AI-content detection.
Triggers on: analyze · score my post · audit this article
Full-site blog health: scores every post, finds orphan pages, cannibalization, stale content and AI-citation gaps via parallel agents.
Triggers on: blog audit · site content health
Detect posts competing for the same keyword/intent by clustering primary keywords from titles and headings.
Triggers on: cannibalization · keyword overlap · competing posts
Verify every statistic by fetching its cited URL and checking the number actually appears on the page.
Triggers on: fact check · verify stats · check sources
Topic-cluster architecture, audience mapping, competitive landscape, AI-citation surface strategy and distribution channels.
Triggers on: blog strategy · content strategy · where to start
Detailed content brief: target keywords, outline, competitive analysis, recommended stats, image/chart suggestions, internal links.
Triggers on: content brief · brief for writer
Editorial calendar with topic clusters, publishing schedule, content-decay detection, freshness plans and seasonal opportunities.
Triggers on: calendar · publishing schedule · plan the month
SERP-based semantic clustering → hub-and-spoke architecture with an internal-link matrix; can auto-execute drafts.
Triggers on: topic cluster · pillar and spokes
Post-writing SEO validation: title length & keyword placement, meta description, heading hierarchy, link audit, canonical.
Triggers on: seo check · validate SEO · pre-publish
Generate complete JSON-LD (BlogPosting, Person, Organization, BreadcrumbList, FAQPage, ImageObject), validated to Google's rules.
Triggers on: schema · structured data · JSON-LD
AI-citation-readiness audit (ChatGPT, Perplexity, Claude, Gemini, AI Overviews) — citability scoring only, no Google changes.
Triggers on: GEO · AI citations · get quoted by AI
FLOW framework for bloggers — evidence-led Find → Optimize → Win loop with 30 stage-specific prompts.
Triggers on: FLOW · evidence-led content
Write → translate → culturally adapt → emit hreflang + sitemap + CMS language map, in one command.
Triggers on: multilingual · write in many languages
Native-quality SEO translation preserving markdown, front-matter, schema, charts and citations.
Triggers on: translate post · localize content
Cultural adaptation after translation — adjusts examples, CTAs, legal references and formality per market.
Triggers on: localize · cultural adaptation
Audit multilingual content for completeness, hreflang correctness, meta parity and freshness; builds a coverage matrix.
Triggers on: locale audit · translation coverage
AI image generation/editing via Gemini — Claude acts as creative director building optimized prompts for hero/section images.
Triggers on: blog image · hero image · generate visual
Audio narration via Gemini TTS — summary, full read-aloud, or two-speaker podcast mode (30 voices) → MP3 + embed.
Triggers on: audio · narration · podcast version
Inline SVG data-viz charts (bar, donut, line, radar…) auto-detecting HTML vs JSX/MDX. Invoked from write/rewrite.
Triggers on: auto — used inside write/rewrite
Google data: PageSpeed, CrUX (25-week), Search Console, URL inspection, Indexing API, GA4 organic, NLP entities, YouTube.
Triggers on: search console · pagespeed · organic data
API-free research of the last 30 days across Reddit, X, YouTube, HN, dev.to, Medium → DISCOURSE.md.
Triggers on: discourse · what are people saying · trends
Query Google NotebookLM notebooks for source-grounded, citation-backed answers from your uploaded documents.
Triggers on: notebooklm · ask my documents
Generate BRAND.md (audience, positioning, editorial do/don'ts, taboo phrases) + VOICE.md — auto-loaded by every sub-skill.
Triggers on: brand · set my voice · house style
Create writing personas using NNGroup's 4 tone dimensions; defines readability targets and vocabulary.
Triggers on: persona · writing voice profile
Extract/suggest/sync tags & categories across WordPress, Shopify, Ghost, Strapi and Sanity.
Triggers on: tags · categories · taxonomy · CMS sync
Turn one post into X/Twitter threads, LinkedIn articles, YouTube scripts, Reddit posts and newsletter excerpts.
Triggers on: repurpose · social posts · newsletter
👷 The specialist agents
The /blog write pipeline is powered by 5 specialist agents that hand off to each other — research, drafting, SEO, scoring and translation.
current 2025–26 statistics, source verification (tier 1–3), royalty-free image discovery
drafting: answer-first formatting, heading hierarchy, sourced stats, natural keyword use
on-page SEO validation: title, meta, headings, internal/external links
the 5-category 100-point score, issue severity, AI-content detection
native-quality translation + localization preserving structure & schema
🔌 Optional data extensions
The core skill works on its own. These plug in third-party data — each needs that provider's own API key. None are required to start.
| Extension | What it adds | Needs |
|---|---|---|
| Gemini (nanobanana MCP) | AI image generation/editing and TTS audio narration for posts. | nanobanana MCP / Gemini |
| Google APIs | PageSpeed, CrUX, Search Console, Indexing, GA4 organic, NLP entities, YouTube data. | Google API credentials |
| NotebookLM | Source-grounded answers from your own uploaded documents. | Google account |
| DataForSEO | SERP/keyword data behind cluster + research (when available). | DataForSEO API key |
| CMS APIs | Tag/category sync + publishing to WordPress, Shopify, Ghost, Strapi, Sanity. | Per-CMS API token |
📄 What the output looks like
Real artifact shapes the skill produces — so you know exactly what you'll get back.
5-category quality score
{
"overall_score": 92,
"verdict": "PASS",
"categories": {
"content_quality": { "score": 94, "notes": "answer-first, sourced" },
"seo_optimization": { "score": 90, "notes": "title+meta+headings ok" },
"eeat_signals": { "score": 91, "notes": "author + citations present" },
"technical_elements": { "score": 95, "notes": "schema + internal links" },
"ai_citation_readiness":{ "score": 90, "notes": "passage-level citable" }
},
"ai_content_detection": { "risk": "low", "burstiness": "human-like" }
}Every /blog write iterates until this clears 90+ with zero criticals (max 3 passes).
The 8 artifacts you get per post
my-post/
├── post.md # the article (front-matter: title, meta, canonical, tags)
├── schema.json # JSON-LD (BlogPosting / FAQPage / ImageObject)
├── meta.md # title + meta description options
├── internal-links.md # suggested internal links
├── images/ # generated hero/section images (if /blog image)
├── audio.mp3 # narration (if /blog audio)
├── score.json # the 5-category score above
└── sources.md # cited sources (fact-checked)One folder, ready to drop into your CMS.
🍳 Recipes — end to end
Copy-paste playbooks for the most common goals. Follow the steps top to bottom.
Publish your first post
Go from a topic to a publish-ready article.
- 1/blog brand initTeach it your voice once (BRAND.md + VOICE.md).
- 2/blog write "your topic"It researches, drafts, fact-checks and scores to 90+.
- 3Review the folder of 8 artifacts and paste
post.mdinto your CMS.
Refresh an old post
The fastest ranking win — update proven content.
- 1/blog analyze ./old.mdScore the current post and see gaps.
- 2/blog rewrite ./old.mdRewrite for Google + AI citations (must beat the original).
- 3/blog seo-check ./old.mdValidate before re-publishing.
Build a topic cluster
Dominate a whole topic, not one keyword.
- 1/blog cluster plan "<seed>"Get a pillar + supporting posts with an internal-link matrix.
- 2/blog cluster executeDraft the cluster.
- 3/blog cannibalization ./contentMake sure none compete with each other.
Go multilingual
Reach global readers without a translation agency.
- 1/blog multilingual "topic" --languages es,fr,deWrite + translate + localize + hreflang in one shot.
- 2/blog locale-audit ./contentVerify coverage, hreflang and freshness.
Weekly content routine
Consistent output that compounds.
- 1/blog calendar monthlyPlan the month's topics.
- 2/blog write "<this week's topic>"Ship one strong post per week.
- 3/blog repurpose ./post.mdTurn it into social + newsletter content.
⌨️ Command explorer
Every command, searchable. Try "audit", "image", "report"…
| Command | Area | What it does |
|---|---|---|
| /blog write <topic> | Create | Write a new post from scratch; iterates to a 90+ quality score. |
| /blog rewrite <file> | Optimize | Optimize an existing post (must score ≥ the original). |
| /blog analyze <file|url> | Audit | Audit blog quality with a 0–100 score. |
| /blog brief <topic> | Plan | Generate a detailed content brief. |
| /blog calendar [monthly|quarterly] | Plan | Generate an editorial calendar. |
| /blog strategy <niche> | Plan | Blog strategy and topic ideation. |
| /blog outline <topic> | Create | SERP-informed outline generation. |
| /blog seo-check <file> | SEO | Post-writing SEO validation. |
| /blog schema <file> | SEO | Generate JSON-LD schema markup. |
| /blog repurpose <file> | Distribute | Repurpose content for other platforms. |
| /blog geo <file> | AI/GEO | AI-citation optimization audit. |
| /blog audit [dir] | Audit | Full-site blog health assessment. |
| /blog image [generate|edit] | Media | AI image generation/editing (Gemini). |
| /blog audio [generate] | Media | Audio narration (Gemini TTS). |
| /blog cannibalization [dir] | Audit | Detect keyword overlap across posts. |
| /blog factcheck <file> | Quality | Verify statistics against cited sources. |
| /blog persona [create|apply] | Voice | Manage writing personas / voice profiles. |
| /blog taxonomy [sync|audit] | CMS | Tag/category CMS management. |
| /blog google [cmd] | Data | Google data: PSI, CrUX, Search Console, GA4, NLP, YouTube, Keywords. |
| /blog cluster [plan|execute] <seed> | Strategy | Semantic topic-cluster planning + execution. |
| /blog multilingual <topic> --languages <codes> | i18n | Write + translate + localize + hreflang. |
| /blog translate <file> --to <codes> | i18n | SEO-optimized translation. |
| /blog localize <file> --locale <code> | i18n | Cultural deep-adaptation per locale. |
| /blog brand [init|show|update] | Voice | Generate BRAND.md + VOICE.md, auto-loaded by all sub-skills. |
| /blog discourse <topic> | Data | API-free last-30-days discourse research. |
📚 Glossary (no jargon left behind)
Every term a non-technical reader might trip on, in one place.
Attracting customers by publishing genuinely useful content instead of (or with) ads.
A schedule of what posts you'll publish and when.
A pillar page plus supporting posts on one theme, to build topical authority.
The metadata block at the top of a Markdown post (title, description, date, tags, canonical).
A tag telling search engines which language/region version of a page to show.
Generative Engine Optimization — making content likely to be cited by AI assistants.
When two of your pages compete for the same keyword and weaken each other.
Structured data describing a page (Article, FAQ, HowTo) to enable rich results.
❓ FAQ
I can't write well. Will the posts sound human?
/blog brand and the skill writes in your tone, fact-checks, and iterates to a high quality score. You review and publish.Does it publish to my site automatically?
Is it free?
How is this different from generic AI writing?
Can it handle other languages?
/blog multilingual writes, translates, localizes and adds hreflang in one command.Claude Blog Graduate
Complete all 15 lessons to unlock your certificate and the Graduate badge.
Awarded by the cabscloud Claude Academy