AgriciDaniel/claude-blog

What is claude‑blog?

claude‑blog is a Claude Code skill suite that lets the Claude AI write, edit, audit, translate, and publish blog posts automatically. It wraps a large set of sub‑skills (31 sub‑skills + an orchestrator) behind a single /blog command set. Each article goes through a 5‑gate Delivery Contract that checks format, visuals, AI‑review scores, link integrity and SEO before the draft is handed back to the user.

---

Who might use it?

• Solo creators who want a turnkey way to produce SEO‑friendly, schema‑rich posts without manual copy‑editing.
• Marketing teams / agencies that need repeatable multi‑post, multilingual, or topic‑cluster workflows.
• Claude Code developers who want a concrete example of a layered Agent Skills architecture (orchestrator → sub‑skills → agents → scripts).

---

Core capabilities (high‑level)

Area	What the skill does
Writing	/blog write <topic> creates a full article (MD, HTML, PDF, hero image, screenshots) from scratch.
Re‑writing	/blog rewrite <file> optimises an existing post for SEO, schema, and AI‑citation readiness.
Planning	/blog strategy, /blog calendar, /blog cluster generate niche strategies, editorial calendars and hub‑and‑spoke topic clusters.
SEO & schema	/blog seo‑check, /blog schema produce Google‑aligned checklists and JSON‑LD markup.
Multilingual	/blog multilingual, /blog translate, /blog localize write, translate and culturally adapt content, emitting proper hreflang tags.
Audio & media	/blog audio (Gemini TTS), /blog image (Gemini image generation) add narration and AI‑generated visuals.
Quality gates	Five automated gates enforce: tool availability, file completeness, visual verification, reviewer score (≥ 90), and asset/link integrity.
Analytics	/blog decay flags traffic drops using Google Search Console data; /blog analyze scores any post 0‑100.
Persona & brand	/blog persona and /blog brand let you store voice profiles and brand guidelines that are automatically loaded by all sub‑skills.
Research	/blog discourse (API‑free social listening), /blog notebooklm (Gemini Notebook research), /blog factcheck verifies cited statistics.

---

Architecture at a glance

• Orchestrator (skills/blog/SKILL.md) parses the /blog command, selects the appropriate sub‑skill, and coordinates agents and helper scripts.
• 31 sub‑skills (skills/blog‑*/SKILL.md) each implement a user‑facing command (e.g., write, rewrite, translate).
• 5 specialized agents (agents/blog‑*.md) perform tasks such as review scoring and citation readiness.
• 22 Python helper scripts (scripts/*.py) implement rendering, pre‑flight checks, quality gates, and installer logic.
• Templates & references provide ready‑made article structures (12 templates) and heuristics used by the delivery contract.
• Test suite (pytest) validates the whole pipeline; CI runs on every push.

---

How to get it

1. Via Claude Code marketplace (recommended):

   /plugin marketplace add AgriciDaniel/claude-blog
   /plugin install claude-blog@agricidaniel-blog
   

   Works on Claude Code 2.1.296 or newer (the skill system was introduced in 2.0.12).
2. Standalone installer (Unix/macOS/Windows) – run the script from the release tag v2.3.0 or later. The installer includes an ownership‑aware cleanup script to avoid accidental deletions.

---

What a typical workflow looks like

/user> /blog strategy tech‑startup
   → generates niche brief, keyword list, and content calendar.
/user> /blog write "How to launch a SaaS MVP"
   → orchestrator runs sub‑skill, creates markdown, renders HTML/PDF, generates hero image.
   → pre‑flight gate 1 checks tools, gate 2 checks file completeness, gate 3 renders screenshots, gate 4 runs the reviewer agent (score ≥ 90), gate 5 verifies links.
   → if any gate fails, the skill iterates (max 3 times) or escalates.
/user> /blog analyze ./output/how-to-launch.md
   → returns a 0‑100 quality score and highlights improvement areas.

All artifacts are saved in an artifact/ folder so you can publish the markdown, HTML, PDF, and images directly to your CMS.

---

Notable extras

• AI Citation Readiness – a heuristic (scripts/ai_citation_score.py) that rates how ready a post is to be cited by other AIs (Perplexity, ChatGPT, etc.).
• Pre‑commit hook – blocks committing a post that scores below 70, ensuring quality even before you run the delivery contract.
• FLOW framework integration – Find, Leverage, Optimize, Win prompts are exposed via /blog flow commands.
• Google API support – PageSpeed, CrUX, Search Console, GA4, NLP, YouTube, Keyword Planner (limited to indexing‑eligible URLs).

---

When to consider it

If you already use Claude Code and want a single‑command, end‑to‑end blog production pipeline that enforces SEO, schema, multilingual publishing, and a rigorous quality gate, claude‑blog provides a ready‑made solution. It is also a valuable learning resource for building complex Claude Code skill suites.