What does this skill do, and when should you use it?
Blog Writer is the claude-blog repository’s skill for creating new blog articles from a topic, brief, or outline. It clarifies audience, search intent, length, and format, selects a content template, presents an outline for approval, and then writes the article. It supports MDX, Markdown, and HTML, with answer-first sections, sourced statistics, citation capsules, internal-linking placeholders, FAQs, images, and SVG charts. It is best suited to users who install the full claude-blog suite because the workflow calls shared agents, references, templates, and scripts.
Accepts a topic, brief, or outline and asks for audience, primary keyword or search intent, word count, and platform when those details are missing. Detects the content type and selects one of 12 templates, then directs research for 8-12 current statistics, images, inline visuals, chart data, and relevant videos. Produces a complete article with frontmatter, a Key Takeaways box, answer-first H2 sections, information-gain markers, citation capsules, internal-linking placeholders, FAQ items, and a conclusion. It can format output as MDX, Markdown, or HTML. Before delivery, the workflow can invoke the repository’s research, chart, image, review, rendering, and preflight components for content, format, visual, asset, and link checks.
- A solo blogger has a topic and wants a structured, research-led first draft instead of a one-shot paragraph response.
- A marketing team needs articles aligned to search intent with sourced statistics, FAQ content, and planned internal-linking zones.
- An SEO editor wants one article workflow aimed at both Google results and AI assistant citation surfaces.
- A team publishing through Next.js MDX, Astro, Hugo, Jekyll, WordPress, Ghost, 11ty, Gatsby, or static HTML needs format-aware article output.
How do you install this skill?
- Fetched web content, external statistics, and project context enter the model instruction environment; indirect prompt injection and source accuracy require human review.
- The core flow depends on Claude Code agents, Python scripts, browser rendering, media platforms, and optional APIs; missing dependencies or mainland-China network restrictions may prevent completion.
- The skill requires many statistics, citations, images, charts, and FAQs but provides no independently verifiable completed output in the supplied file set; delivery-gate scores are not a substitute for factual verification.
- Version references differ between v1.9.0 and v1.9.1 across the supplied materials; verify the actual revision and dependency set before installation.
- Shell / CLI
- Network access
- Local filesystem
- MCP Server
Python 3.11+Claude Codeblog-researcher agentblog-reviewer agent
The skill is located at skills/blog-write/SKILL.md and is normally installed as part of the full claude-blog suite. A documented repository installation is: git clone https://github.com/AgriciDaniel/claude-blog.git; cd claude-blog; git checkout v1.9.0; chmod +x install.sh && ./install.sh. The README does not document a standalone blog-write installation path. Restart Claude Code after installation.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/AgriciDaniel/claude-blog.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/blog-write" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- write blog about [topic]
- create article about [topic]
- new blog post,
- generate blog post
In Claude Code, use a request such as “write blog about [topic]” or “create article about [topic]”; the skill description also lists “new blog post,” “draft blog,” and “generate blog post” as triggers. If you provide only a topic, be ready to specify the audience, primary keyword or search intent, desired word count, and output format. The workflow presents an outline for approval before writing and relies on the repository’s supporting agents and scripts for the full delivery process.
What are this skill's strengths and limitations?
- Covers topic clarification, template selection, research planning, outlining, writing, and quality checks in one workflow.
- Defines concrete structures for answer-first writing, sourced statistics, citation capsules, FAQs, internal links, and originality markers.
- Supports MDX, Markdown, and HTML and specifies image, video, and SVG chart handling.
- Uses a blocking delivery review requiring a score of at least 90/100 and zero P0 issues.
- It is not a standalone single-file tool; it references shared agents, templates, reference documents, and Python scripts in the repository.
- Research, statistics, image, and video discovery require network access, while some image-generation paths depend on optional MCP configuration.
- The outline-approval step adds interaction and may be unnecessary for a quick throwaway draft.
- The supplied material does not establish that every external platform, media API, or optional tool will be configured in every environment.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Blog Writer this page | 44 · Not recommended | ★ 2.4k | 14d ago | MIT |
| Blog Rewrite Optimizer | 45 · Not recommended | ★ 2.4k | 14d ago | MIT |
| Claude Blog Content Engine | 47 · Use with care | ★ 2.4k | 14d ago | MIT |
| Blog Brief Planner | 44 · Not recommended | ★ 2.4k | 14d ago | MIT |
| Blog Voice Learner | 46 · Use with care | ★ 2.4k | 14d ago | MIT |
Compared with prompting Claude or ChatGPT directly, Blog Writer adds template selection, research planning, citation structure, and delivery checks. Compared with hosted tools such as Copy.ai or Jasper, it is an open-source Claude Code workflow. Compared with building a custom pipeline, it offers predefined editorial rules and quality gates but requires more repository components and setup.
How did FollowSkills review this skill?
The skill requires web search, WebFetch, curl, external image/video embedding, scripts, and agent calls. It requires source names, URLs, and retrieval dates, and asks for outline approval, but it does not clearly define least privilege, sensitive-data handling, external-content warnings, write scope, rollback, or final delivery confirmation. No explicit malware, credential theft, or destructive default is shown.
The workflow and blocking gates are fairly structured, with template fallback, failure diagnostics, and a maximum of three iterations. However, the selected file does not verify availability of its agents, references, scripts, or external services; several features fail silently, and abnormal-input and failure-message behavior is thin. Static calibration limits this score to 10.
Triggers, topic inputs, output formats, and supported blog platforms are clearly described, and some non-fit cases are stated. However, the workflow is primarily optimized for English SEO and overseas search, media, and video services. Chinese-language workflows, mainland-China reachability, and precise non-trigger boundaries are not established.
Frontmatter, argument hint, phased structure, template mapping, examples, quality checks, MIT licensing, and a versioned delivery contract are documented. Nevertheless, the skill depends heavily on other skills, references, scripts, and agents not defined in this file, and it lacks self-contained installation, compatibility, troubleshooting, changelog, and maintenance-owner guidance.
The workflow covers topic clarification, research, drafting, media, rendering, review, and delivery, with directly usable Markdown, MDX, or HTML structures. Yet substantial manual verification remains for sources, links, assets, and placeholders, and static evidence cannot confirm reliable end-to-end delivery. The static ceiling therefore limits this to 7.
The skill specifies source records, citation formatting, quality gates, and audit artifacts, while repository context indicates CI and tests. The supplied tests mainly cover analysis utilities, not blog-write’s core path, external research, rendering, media retrieval, agent dispatch, or the five gates. Static evidence supports only limited reproducibility.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →