[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7BKNY1WH8MHVGCBPC6SR5SH":3,"skill-same:01M7BKNY1WH8MHVGCBPC6SR5SH":48},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":6,"ownerHtmlUrl":32,"ownerBlogUrl":6,"ownerDescription":33,"updatedAt":34,"install":35,"resources":47},"01M7BKNY1WH8MHVGCBPC6SR5SH","blog-write",null,"Write new blog articles from scratch optimized for Google rankings and AI citations. Generates full articles with template selection, answer-first formatting, Key Takeaways summary box, information gain markers, citation capsules, sourced statistics, Pixabay\u002FUnsplash images, built-in SVG chart generation, FAQ schema, internal linking zones, and proper heading hierarchy. Supports MDX, markdown, and HTML output. Use when user says \"write blog\", \"new blog post\", \"create article\", \"write about\", \"draft blog\", \"generate blog post\".","om-scogo\u002Fskillsh-scraper","om-scogo","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F160466840?v=4","01M79AWB6DFVCQ7RK70CS34B96","Om Achrekar","01M79AWB6ZM9SKR8KDH1WSSM22",0,"MIT","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fagricidaniel-claude-blog\u002Fblog-write\u002Fskills\u002Fblog-write\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: blog-write\ndescription: >\n  Write new blog articles from scratch optimized for Google rankings and AI\n  citations. Generates full articles with template selection, answer-first\n  formatting, Key Takeaways summary box, information gain markers, citation capsules, sourced\n  statistics, Pixabay\u002FUnsplash images, built-in SVG chart generation, FAQ schema,\n  internal linking zones, and proper heading hierarchy. Supports MDX, markdown,\n  and HTML output.\n  Use when user says \"write blog\", \"new blog post\", \"create article\",\n  \"write about\", \"draft blog\", \"generate blog post\".\nuser-invokable: true\nargument-hint: \"\u003Ctopic>\"\nlicense: MIT\n---\n\n# Blog Writer: New Article Generation\n\nWrites complete blog articles from a topic, brief, or outline. Every article\nfollows the 6 pillars of dual optimization (Google rankings + AI citations).\n\n**Key references** (paths relative to repo root; references live in the\nmain `blog` skill's references directory, not in `blog-write\u002F`):\n\n- `skills\u002Fblog\u002Freferences\u002Fsynthesis-contract.md`: 6 LAWs for synthesis output (v1.8.0; applies whenever the article embeds research-synthesis prose)\n- `skills\u002Fblog\u002Freferences\u002Fcontent-templates.md`: Template selection guide and usage\n- `skills\u002Fblog\u002Freferences\u002Fquality-scoring.md`: 5-category scoring (Content 30, SEO 25, E-E-A-T 15, Technical 15, AI Citation 15)\n- `skills\u002Fblog\u002Freferences\u002Feeat-signals.md`: Experience, expertise, authority, trust markers\n- `skills\u002Fblog\u002Freferences\u002Finternal-linking.md`: Linking strategy and anchor text rules\n- `skills\u002Fblog\u002Freferences\u002Fvisual-media.md`: Image sourcing and chart styling\n\n## Workflow\n\n### Phase 0: Surface Targeting (do this BEFORE research)\n\nDecide which of the FLOW 5 surfaces this post is meant to win. The choice\nshapes structure, length, citation density, and call-to-action. The 5 surfaces\nin 2026:\n\n1. Owned site (organic Google ranking)\n2. SERP including AI Overviews\n3. AI assistant citations (ChatGPT, Perplexity, Claude, Gemini, Copilot, You.com)\n4. Local pack (out of scope for blog content; use claude-seo for local)\n5. Communities and video (Reddit, YouTube, LinkedIn, Quora, niche forums)\n\nMost posts target surfaces 1, 2, and 3 by default. If the same query also\nsurfaces in a community (Reddit thread, YouTube comment), apply dual-surface\nthinking: optimize the post for extraction AND plan a community echo (covered\nin `\u002Fblog repurpose`).\n\nFor a deeper surface-by-surface workflow, see\n`skills\u002Fblog\u002Freferences\u002Fflow-alignment.md` and `\u002Fblog flow find`.\n\n### Phase 1: Topic Understanding\n\n1. **Clarify the topic** - If the user provides just a topic, ask:\n   - Target audience (who is this for?)\n   - Primary keyword \u002F search intent\n   - Desired word count (default: 2,000-2,500 words)\n   - Platform\u002Fformat (MDX, markdown, HTML - auto-detect if in a project)\n2. **If a brief exists** - Load it and skip to Phase 1.5\n\n### Phase 1.5: Template Selection\n\nSelect the appropriate content template from the 12 templates in\n`skills\u002Fblog\u002Ftemplates\u002F` (the main `blog` skill owns the templates directory).\n\n1. **Auto-detect content type** from the topic and search intent:\n   | Signal | Template |\n   |--------|----------|\n   | \"How to...\", process, steps | `how-to-guide` |\n   | \"Best X\", \"Top N\", list format | `listicle` |\n   | Client result, before\u002Fafter, metrics | `case-study` |\n   | \"X vs Y\", comparison, alternatives | `comparison` |\n   | Broad topic, comprehensive guide | `pillar-page` |\n   | \"Is X worth it\", product evaluation | `product-review` |\n   | Opinion, prediction, industry take | `thought-leadership` |\n   | Expert quotes, multi-source collection | `roundup` |\n   | Code walkthrough, tool demo, technical | `tutorial` |\n   | Breaking news, algorithm update, event | `news-analysis` |\n   | Survey results, experiment, original data | `data-research` |\n   | Q&A, knowledge base, \"What is X\" | `faq-knowledge` |\n\n2. **Load the matching template**: Read from `skills\u002Fblog\u002Ftemplates\u002F\u003Ctype>.md`\n3. **Adapt the outline** - Use the template's section structure, heading patterns,\n   and word count guidance to shape Phase 3's outline\n4. **Fallback** - If no template clearly fits, use the generic outline structure\n   in Phase 3 below. Inform the user which template was selected (or that none matched).\n\nSee `skills\u002Fblog\u002Freferences\u002Fcontent-templates.md` for detailed selection criteria and intent mapping.\n\n### Phase 2: Research\n\nSpawn a `blog-researcher` agent (or do inline research with WebSearch):\n\n1. **Find 8-12 current statistics** (2025-2026 data preferred)\n   - Search: `[topic] study 2025 2026 data statistics`\n   - Prioritize tier 1-3 sources (see `skills\u002Fblog\u002Freferences\u002Fquality-scoring.md`)\n   - Record: statistic, source name, URL, date, methodology\n2. **Find a cover image** (wide, high-quality, topic-relevant):\n   - Search: `site:pixabay.com [topic] wide banner` (preferred)\n   - Alternative: `site:unsplash.com [topic] wide`\n   - Fallback: `site:pexels.com [topic] wide banner`\n   - Target dimensions: 1200x630 (OG-compatible) or 1920x1080\n   - Or generate a custom SVG cover via `blog-chart` (text-on-gradient with key stat)\n   - Or generate a custom AI image via `blog-image` sub-skill (if nanobanana-mcp configured)\n   - See `skills\u002Fblog\u002Freferences\u002Fvisual-media.md` for cover image sizing details\n3. **Find 3-5 inline images** from open-source platforms:\n   - **Pixabay** (preferred): Search `site:pixabay.com [topic keywords]`\n     - Extract image URL from page\n     - Direct URLs: `https:\u002F\u002Fcdn.pixabay.com\u002Fphoto\u002FYYYY\u002FMM\u002FDD\u002FHH\u002FMM\u002Ffilename.jpg`\n     - Verify with `curl -sI \"\u003Curl>\" | head -1` returns HTTP 200\n   - **Unsplash** (alternative): Search `site:unsplash.com [topic keywords]`\n     - Build URL: `https:\u002F\u002Fimages.unsplash.com\u002Fphoto-\u003Cid>?w=1200&h=630&fit=crop&q=80`\n   - **Pexels** (fallback): Search `site:pexels.com [topic keywords]`\n4. **Plan 2-4 data visualizations** from researched statistics\n   - Select diverse chart types (see `skills\u002Fblog\u002Freferences\u002Fvisual-media.md`)\n   - Map data points to chart formats\n5. **AI image generation** (optional, if nanobanana-mcp configured):\n   - If stock photo results are insufficient (\u003C 3 good matches) or topic is too niche\n   - Generate custom hero image and\u002For inline illustrations via `blog-image` sub-skill\n   - Stock photos remain default - AI generation supplements, never replaces\n6. **NotebookLM research** (optional, if user has relevant notebooks):\n   - If the user mentions a NotebookLM notebook or the topic aligns with a configured notebook\n   - Query via `blog-notebooklm` for source-grounded data from user-uploaded documents\n   - Treat NotebookLM responses as Tier 1 sources (user's own primary documents)\n   - Falls back silently if not configured or not authenticated\n7. **Find relevant YouTube videos** (2-3 per post):\n   - Use `blog-google` youtube command or WebSearch `site:youtube.com [topic] [year]`\n   - Apply quality criteria from `skills\u002Fblog\u002Freferences\u002Fvideo-embeds.md` (min score 50\u002F100)\n   - Select 2-3 best videos. Falls back silently if none found.\n\n### Phase 3: Outline Generation\n\nCreate a structured outline before writing. If a template was loaded in Phase 1.5,\nadapt this skeleton to match the template's section structure:\n\n```\n# [Title as Question - Include Primary Keyword]\n\n## Introduction (100-150 words)\n- Hook with surprising statistic\n- Problem\u002Fopportunity statement\n- What the reader will learn\n\n> **Key Takeaways**\n> - [Core finding with statistic and source]\n> - [Second key insight or recommendation]\n> - [Third actionable takeaway]\n> (3-5 bullets, 40-60 words combined)\n\n## H2: [Question Format] (300-400 words)\n- Answer-first paragraph (40-60 words with stat + source)\n- Supporting evidence\n- [Image placement]\n- Practical advice\n- [CITATION CAPSULE placeholder]\n- [INTERNAL-LINK: anchor text → target description]\n\n## H2: [Question Format] (300-400 words)\n- Answer-first paragraph\n- [Chart: type + data description]\n- Analysis and implications\n- [CITATION CAPSULE placeholder]\n- [INTERNAL-LINK: anchor text → target description]\n\n## H2: [Statement for Variety] (300-400 words)\n- Answer-first paragraph\n- Real-world example or case study\n- [Image placement]\n- [CITATION CAPSULE placeholder]\n\n## H2: [Question Format] (300-400 words)\n- Answer-first paragraph\n- [Chart: type + data description]\n- Step-by-step guidance\n- [CITATION CAPSULE placeholder]\n- [INTERNAL-LINK: anchor text → target description]\n\n## H2: [Question Format] (200-300 words)\n- Answer-first paragraph\n- Forward-looking analysis\n\n## [CTA Section or Inline Placement]\n- See `skills\u002Fblog\u002Freferences\u002Fcta-placement.md` for placement rules by content type\n- Place CTA after value delivery, not at arbitrary positions\n- Single focused CTA per post (266% more conversions)\n- [CTA: contextual call-to-action matching article topic]\n\n## FAQ Section (3-5 questions, 40-60 words each answer)\n- [INTERNAL-LINK: anchor text → detailed content]\n\n## Conclusion (100-150 words)\n- Key takeaways (bulleted)\n- Call to action\n- [INTERNAL-LINK: anchor text → next logical content]\n```\n\nPresent the outline to the user for approval before writing.\n\n**Visual element pacing**: Insert `[IMAGE]`, `[CHART]`, `[VIDEO]`, or `[CALLOUT]` markers\nevery 300-500 words. Alternate types (no consecutive same-type). See\n`skills\u002Fblog\u002Freferences\u002Fcontent-rules.md` Visual Rhythm section and\n`skills\u002Fblog\u002Freferences\u002Fcta-placement.md` for CTA positioning.\n\n### Phase 4: Chart Generation (Built-In)\n\nWhen the researcher identifies chart-worthy data (3+ comparable metrics, trend data,\nbefore\u002Fafter comparisons):\n\n1. Select chart type using the diversity rule (no repeated types per post)\n2. Invoke `blog-chart` sub-skill with: chart type, title, data values, source, platform format\n3. Embed the returned SVG directly in the post within a `\u003Cfigure>` wrapper\n4. Target 2-4 charts per 2,000-word post\n5. Distribute charts evenly - never cluster them\n\nSee `skills\u002Fblog\u002Freferences\u002Fvisual-media.md` for chart type selection and styling rules.\n\n### Phase 5: Content Writing\n\nWrite the full article following these rules:\n\n#### 5a. Frontmatter\n```yaml\n---\ntitle: \"[Question-format title with primary keyword]\"\ndescription: \"[Fact-dense, 150-160 chars, includes 1 statistic]\"\ncoverImage: \"[URL from Pixabay\u002FUnsplash\u002FPexels or generated SVG path]\"\ncoverImageAlt: \"[Descriptive sentence about the cover image]\"\nogImage: \"[Same as coverImage, or custom OG image URL]\"\ndate: \"YYYY-MM-DD\"\nlastUpdated: \"YYYY-MM-DD\"\nauthor: \"[Author name]\"\ntags: [\"keyword1\", \"keyword2\", \"keyword3\"]\n---\n```\n\nIf the platform uses a different field name (e.g., `image`, `hero`, `thumbnail`),\nadapt to match the project's existing frontmatter convention.\n\n#### 5b. Summary Box (Key Takeaways)\n\nImmediately after the introduction (before the first H2 body section), add a summary box:\n\n```markdown\n> **Key Takeaways**\n> - [Core finding with statistic] ([Source], year)\n> - [Second key insight or recommendation]\n> - [Third actionable takeaway]\n```\n\nRequirements:\n- 3-5 bullet points, 40-60 words combined\n- Must be self-contained - understandable without reading the article\n- Include 1 specific statistic with source name\n- State the key finding, recommendation, or answer\n- Default label: \"Key Takeaways\". If a persona is active, use the persona's summary_label\n- Backward compatible: accept existing TL;DR boxes during rewrites\n\n#### 5c. Answer-First Formatting (Critical)\nEvery H2 section MUST open with a 40-60 word paragraph containing:\n- At least one specific statistic with source attribution\n- A direct answer to the heading's implicit question\n\nPattern:\n```markdown\n## How Does X Impact Y in 2026?\n\n[Stat from source] ([Source Name](url), year). [Direct answer to the heading\nquestion in 1-2 more sentences, explaining the implication and what this means\nfor the reader.]\n```\n\n**FLOW evidence triple (drafting requirement, not just audit):**\n\nEvery public statistic must carry three components AT DRAFTING TIME:\n\n1. **Year anchor in prose.** Write \"In 2026,\" or \"As of Q1 2026,\" BEFORE\n   the statistic, in the sentence body. Year buried inside parentheses\n   does not count. Example:\n   - GOOD: \"In 2026, Ahrefs found a 58% lower CTR for position one when\n     an AI Overview was present.\"\n   - WEAK: \"Position-one CTR dropped 58% (Ahrefs, 2026).\"\n\n2. **Inline citation with publisher and title.** Name both the publisher\n   and the document title (or report name), not just a brand. Example:\n   - GOOD: \"Ahrefs, AI Overviews CTR update, December 2025\"\n   - WEAK: \"Ahrefs reported...\"\n\n3. **URL plus retrieval date in the source block at the bottom of the post.**\n   Provenance discipline lets future readers and AI crawlers verify the\n   source still says what was claimed. Format:\n   - \"[Publisher], [Title], retrieved YYYY-MM-DD, [full URL]\"\n\n**FLOW quality bar (drop or replace):**\nPublic claims must use verified sources OR stay qualitative. If a statistic\ncannot be verified, drop it. If it is contradicted by a more recent source,\nreplace it with the verified alternative. Do not soften vague language to\nkeep an unsourceable number.\n\nFor evidence-led optimization prompts (CTR audit, AI detector test, schema,\nPAA rewording, ChatGPT visibility), see `\u002Fblog flow optimize`.\n\n#### 5d. Information Gain Markers\n\nDistribute at least 2-3 information gain markers throughout the article. These\nsignal to search engines and AI systems that the content contains original value\nnot available elsewhere.\n\nTag each with a comment or visible marker:\n\n- `[ORIGINAL DATA]` - Proprietary surveys, experiments, A\u002FB test results, case\n  study metrics the author collected first-hand\n- `[PERSONAL EXPERIENCE]` - First-hand observations, lessons learned from direct\n  involvement, \"when we tried X, Y happened\" narratives\n- `[UNIQUE INSIGHT]` - Analysis others haven't made, contrarian perspectives\n  backed by data, novel connections between existing research\n\nPlacement:\n- Weave into the body text naturally\n- Use as inline comments: `\u003C!-- [ORIGINAL DATA] -->` before the relevant paragraph\n- Or as visible callouts if the format supports it:\n  ```markdown\n  > **Our finding:** [original observation backed by specific data]\n  ```\n- Minimum 2 per post, target 3 for comprehensive articles\n\nThese markers map directly to the \"Originality\u002Funique value markers\" criterion\nin the Content Quality scoring category (see `skills\u002Fblog\u002Freferences\u002Fquality-scoring.md`).\n\n#### 5e. Citation Capsules\n\nFor each major H2 section, generate a citation capsule - a 40-60 word self-contained\npassage designed so AI systems can extract and quote it directly.\n\nRequirements per capsule:\n- 40-60 words, self-contained (makes sense in isolation)\n- Contains: one specific claim + one data point + source attribution\n- Written in a declarative, quotable style\n- Placed within the H2 section body (not as a separate block)\n\nExample:\n```markdown\nAccording to a 2026 Gartner study, 58% of enterprise buyers now consult AI\nassistants before contacting a vendor ([Gartner](https:\u002F\u002Fwww.gartner.com), 2026).\nThis shift means B2B content must answer specific questions concisely enough\nfor AI systems to extract and cite in their responses.\n```\n\nCapsules map to the \"AI Citation Readiness\" scoring category (15 points) in\n`skills\u002Fblog\u002Freferences\u002Fquality-scoring.md`.\n\n#### 5f. Internal Linking Zones\n\nMark internal linking opportunities throughout the article using placeholder\nnotation. The user (or a follow-up pass) will resolve these to actual URLs.\n\nZone placement:\n- **Introduction** - Link to related pillar content or topic hub\n- **Each H2 section** - Link to supporting articles, deeper dives, related tools\n- **FAQ section** - Link answers to detailed content that expands on the answer\n- **Conclusion** - Link to the next logical piece of content the reader should consume\n\nFormat:\n```markdown\n[INTERNAL-LINK: anchor text → target description]\n```\n\nExample:\n```markdown\nFor a deeper dive into keyword clustering, see our\n[INTERNAL-LINK: complete guide to keyword clustering → pillar page on keyword research methodology].\n```\n\nTarget 5-10 internal link zones per 2,000-word post. Use descriptive anchor text\n(never \"click here\" or \"read more\"). See `skills\u002Fblog\u002Freferences\u002Finternal-linking.md` for\nanchor text rules and linking strategy.\n\n#### 5g. Paragraph Rules\n- Every paragraph: 40-80 words (never exceed 150)\n- Every sentence: max 15-20 words\n- Start each paragraph with the most important information\n- Target Flesch Reading Ease: 60-70\n\n#### 5h. Heading Rules\n- One H1 (title only)\n- H2s for main sections (60-70% as questions)\n- H3s for subsections only - never skip levels\n- Include primary keyword naturally in 2-3 headings\n\n#### 5i. Image Embedding\n\nStandard markdown:\n```markdown\n![Descriptive alt text - topic keywords naturally](https:\u002F\u002Fcdn.pixabay.com\u002Fphoto\u002F...)\n```\n\nMDX with Next.js Image (if detected):\n```mdx\n![Descriptive alt text - topic keywords naturally](https:\u002F\u002Fcdn.pixabay.com\u002Fphoto\u002F...)\n```\n\n- Place images after H2 headings, before body text\n- Space evenly throughout the post (not clustered)\n- Alt text should be a full descriptive sentence\n\n#### 5j. Chart Embedding\n\nStandard markdown\u002FHTML:\n```html\n\u003Cfigure>\n  \u003Csvg viewBox=\"0 0 560 380\" ...>...\u003C\u002Fsvg>\n  \u003Cfigcaption>Source: [Source Name], [Year]\u003C\u002Ffigcaption>\n\u003C\u002Ffigure>\n```\n\nMDX format:\n```mdx\n\u003Cfigure className=\"chart-container\" style={{margin: '2.5rem 0', textAlign: 'center', padding: '1.5rem', borderRadius: '12px'}}>\n  \u003Csvg viewBox=\"0 0 560 380\" ...>...\u003C\u002Fsvg>\n\u003C\u002Ffigure>\n```\n\n#### 5k. Video Embedding\nEmbed YouTube videos using srcdoc lazy-loading pattern from `skills\u002Fblog\u002Freferences\u002Fvideo-embeds.md`.\nInclude aria-label, noscript fallback for AI crawlers. Place after relevant H2, 500+ words apart.\n\n#### 5l. Citation Format\nInline attribution (always):\n```markdown\nOrganic CTR declined 61% with AI Overviews ([Seer Interactive](https:\u002F\u002Fwww.seerinteractive.com\u002F), 2025).\n```\n\n#### 5m. FAQ Section\nAdd 3-5 FAQ items with 40-60 word answers. Each answer must contain a statistic.\n\nFor MDX with FAQSchema component:\n```mdx\n\u003CFAQSchema faqs={[\n  { question: \"Question?\", answer: \"40-60 word answer with statistic and source.\" },\n]} \u002F>\n```\n\nFor standard markdown:\n```markdown\n## Frequently Asked Questions\n\n### Question text here?\n\nAnswer with statistic and source attribution (40-60 words).\n```\n\n#### 5n. Internal Linking\n- 5-10 internal links per 2,000-word post\n- Link to relevant existing content naturally\n- Use descriptive anchor text (not \"click here\")\n\n### Phase 6: Quality Check\n\nBefore delivering, verify:\n\n#### Structure and Content\n1. Every H2 opens with a statistic + source\n2. No paragraph exceeds 150 words\n3. All statistics have named tier 1-3 sources\n4. 2-4 charts with type diversity\n5. 3-5 inline images with descriptive alt text\n6. Cover image present in frontmatter (coverImage + ogImage)\n7. FAQ section present with 3-5 items\n8. Heading hierarchy is clean (H1 -> H2 -> H3)\n9. Meta description is 150-160 chars with a stat\n\n#### New Element Verification\n10. TL;DR box present after introduction (40-60 words, contains statistic + source)\n11. At least 2-3 information gain markers (`[ORIGINAL DATA]`, `[PERSONAL EXPERIENCE]`, or `[UNIQUE INSIGHT]`)\n12. Citation capsules present in major H2 sections (40-60 words, self-contained, quotable)\n13. Internal linking zones marked in introduction, H2 sections, FAQ, and conclusion\n14. No AI-detectable phrases from banned list (see `agents\u002Fblog-writer.md`)\n\n#### Burstiness and Naturalness Check\n15. **Sentence length variance** - Verify a mix of short (8-word) and long (25-word) sentences. Uniform sentence length signals AI authorship.\n16. **Banned AI phrase scan** - Check for and remove:\n    - \"in today's digital landscape\", \"it's important to note\", \"dive into\"\n    - \"game-changer\", \"navigate the landscape\", \"revolutionize\", \"seamlessly\"\n    - \"cutting-edge\", \"harness the power of\", \"leverage\" (as verb)\n    - \"delve\", \"crucial\", \"elevate\", \"foster\", \"landscape\" (overused)\n    - \"multifaceted\", \"robust\", \"tapestry\", \"embark\"\n    - Full list in `agents\u002Fblog-writer.md`\n17. **Contractions** - Verify natural use of contractions (\"it's\", \"we've\", \"don't\", \"isn't\"). Formal AI prose avoids contractions; natural writing uses them.\n18. **Rhetorical questions** - Verify at least one rhetorical question every 200-300 words to break up declarative patterns.\n19. **YouTube videos** - 2-3 embeds with lazy loading, aria-labels, and noscript fallback (see `skills\u002Fblog\u002Freferences\u002Fvideo-embeds.md`)\n\n### Phase 6.5: Delivery Contract Enforcement (v1.9.0)\n\nBefore Phase 7, run the 5-gate delivery contract per `skills\u002Fblog\u002Freferences\u002Fblog-delivery-contract.md`. The user is never the first reviewer; the gates are.\n\nSteps:\n\n1. **Capability discovery + hero**: run `python scripts\u002Fblog_preflight.py --draft \u003Cfolder> --gate 1` to enumerate available paths. If `nanobanana-mcp` is loaded, generate the hero via the MCP tool. Otherwise run `python scripts\u002Fgenerate_hero.py --topic \"\u003Ctitle>\" --tags \"\u003Ctags>\" --out \u003Cfolder>` (uses the Gemini, Unsplash, Pexels, Pixabay, Openverse ladder).\n\n2. **Format completeness**: render the canonical `.md` to `.html` and `.pdf` via `python scripts\u002Fblog_render.py --md \u003Cslug>.md --out-dir \u003Cfolder>`. All three artifacts plus `hero.\u003Cext>` must end up in the draft folder.\n\n3. **Content review (blocking)**: dispatch the `blog-reviewer` agent (Task tool) against the rendered `.html`. The agent emits its scorecard to `\u003Cfolder>\u002Freview.md` ending with `BLOCKING: true|false (reason)`. Threshold: overall score 90\u002F100 or higher AND zero P0 issues per `editorial-heuristics.md`.\n\n4. **Visual + asset gates**: run `python scripts\u002Fblog_preflight.py --draft \u003Cfolder> --strict`. This runs Gate 3 (visual verification via patchright at 3 viewport widths), Gate 4 (reads review.md BLOCKING line), and Gate 5 (asset + link integrity). Exit 0 = ship; exit 1 = block.\n\n5. **Iteration**: on any block, capture the failure diagnostic from `\u003Cfolder>\u002Fpreflight-report.json`, re-dispatch the blog-writer agent with the diagnostic as input, and re-run from step 1. Maximum 3 iterations. On the 3rd failure, STOP and present the failure diagnostic instead of the draft.\n\nThe orchestrator holds the loop counter; this sub-skill never loops itself.\n\n### Phase 7: Delivery\n\nPresent the completed article ONLY after Phase 6.5 returns all gates passing. Include the screenshots from `\u003Cfolder>\u002Fpreview\u002F*.png` in the summary so the user can see what they are getting before reading the prose.\n\nSummary template:\n\n```\n## Blog Post Complete: [Title]\n\n### Template Used\n- [Template name] (or \"generic outline - no template matched\")\n\n### Statistics\n- [N] sourced statistics from tier 1-3 sources\n- [N] unique sources cited\n\n### Visual Elements\n- Cover image: [source - Pixabay\u002FUnsplash\u002FPexels or generated SVG]\n- [N] inline images (Pixabay\u002FUnsplash\u002FPexels)\n- [N] SVG charts (types: bar, lollipop, donut, line)\n- [N] YouTube video embeds (titles: ...)\n\n### Dual-Optimization Elements\n- TL;DR box: present (N words)\n- Information gain markers: [N] ([types used])\n- Citation capsules: [N] across H2 sections\n- Internal linking zones: [N] marked\n\n### Structure\n- [N] H2 sections with answer-first formatting\n- [N] FAQ items with schema\n- Word count: ~[N] words\n- Estimated reading time: [N] min\n\n### Naturalness\n- Sentence length variance: [pass\u002Ffail]\n- AI phrase scan: [pass\u002Ffail]\n- Contractions used: [yes\u002Fno]\n- Rhetorical questions: [N] (target: 1 per 200-300 words)\n\n### Next Steps\n- Review and customize for your brand voice\n- Resolve [INTERNAL-LINK] placeholders with actual URLs\n- Add internal links to your existing content\n- Run `\u002Fblog analyze \u003Cfile>` to verify quality score\n- Generate VideoObject schema: `\u002Fblog schema \u003Cfile>` (includes video markup)\n- Generate audio narration: `\u002Fblog audio generate \u003Cfile>` (optional)\n```\n","# Blog Writer: New Article Generation\n\nWrites complete blog articles from a topic, brief, or outline. Every article\nfollows the 6 pillars of dual optimization (Google rankings + AI citations).\n\n**Key references** (paths relative to repo root; references live in the\nmain `blog` skill's references directory, not in `blog-write\u002F`):\n\n- `skills\u002Fblog\u002Freferences\u002Fsynthesis-contract.md`: 6 LAWs for synthesis output (v1.8.0; applies whenever the article embeds research-synthesis prose)\n- `skills\u002Fblog\u002Freferences\u002Fcontent-templates.md`: Template selection guide and usage\n- `skills\u002Fblog\u002Freferences\u002Fquality-scoring.md`: 5-category scoring (Content 30, SEO 25, E-E-A-T 15, Technical 15, AI Citation 15)\n- `skills\u002Fblog\u002Freferences\u002Feeat-signals.md`: Experience, expertise, authority, trust markers\n- `skills\u002Fblog\u002Freferences\u002Finternal-linking.md`: Linking strategy and anchor text rules\n- `skills\u002Fblog\u002Freferences\u002Fvisual-media.md`: Image sourcing and chart styling\n\n## Workflow\n\n### Phase 0: Surface Targeting (do this BEFORE research)\n\nDecide which of the FLOW 5 surfaces this post is meant to win. The choice\nshapes structure, length, citation density, and call-to-action. The 5 surfaces\nin 2026:\n\n1. Owned site (organic Google ranking)\n2. SERP including AI Overviews\n3. AI assistant citations (ChatGPT, Perplexity, Claude, Gemini, Copilot, You.com)\n4. Local pack (out of scope for blog content; use claude-seo for local)\n5. Communities and video (Reddit, YouTube, LinkedIn, Quora, niche forums)\n\nMost posts target surfaces 1, 2, and 3 by default. If the same query also\nsurfaces in a community (Reddit thread, YouTube comment), apply dual-surface\nthinking: optimize the post for extraction AND plan a community echo (covered\nin `\u002Fblog repurpose`).\n\nFor a deeper surface-by-surface workflow, see\n`skills\u002Fblog\u002Freferences\u002Fflow-alignment.md` and `\u002Fblog flow find`.\n\n### Phase 1: Topic Understanding\n\n1. **Clarify the topic** - If the user provides just a topic, ask:\n   - Target audience (who is this for?)\n   - Primary keyword \u002F search intent\n   - Desired word count (default: 2,000-2,500 words)\n   - Platform\u002Fformat (MDX, markdown, HTML - auto-detect if in a project)\n2. **If a brief exists** - Load it and skip to Phase 1.5\n\n### Phase 1.5: Template Selection\n\nSelect the appropriate content template from the 12 templates in\n`skills\u002Fblog\u002Ftemplates\u002F` (the main `blog` skill owns the templates directory).\n\n1. **Auto-detect content type** from the topic and search intent:\n   | Signal | Template |\n   |--------|----------|\n   | \"How to...\", process, steps | `how-to-guide` |\n   | \"Best X\", \"Top N\", list format | `listicle` |\n   | Client result, before\u002Fafter, metrics | `case-study` |\n   | \"X vs Y\", comparison, alternatives | `comparison` |\n   | Broad topic, comprehensive guide | `pillar-page` |\n   | \"Is X worth it\", product evaluation | `product-review` |\n   | Opinion, prediction, industry take | `thought-leadership` |\n   | Expert quotes, multi-source collection | `roundup` |\n   | Code walkthrough, tool demo, technical | `tutorial` |\n   | Breaking news, algorithm update, event | `news-analysis` |\n   | Survey results, experiment, original data | `data-research` |\n   | Q&A, knowledge base, \"What is X\" | `faq-knowledge` |\n\n2. **Load the matching template**: Read from `skills\u002Fblog\u002Ftemplates\u002F\u003Ctype>.md`\n3. **Adapt the outline** - Use the template's section structure, heading patterns,\n   and word count guidance to shape Phase 3's outline\n4. **Fallback** - If no template clearly fits, use the generic outline structure\n   in Phase 3 below. Inform the user which template was selected (or that none matched).\n\nSee `skills\u002Fblog\u002Freferences\u002Fcontent-templates.md` for detailed selection criteria and intent mapping.\n\n### Phase 2: Research\n\nSpawn a `blog-researcher` agent (or do inline research with WebSearch):\n\n1. **Find 8-12 current statistics** (2025-2026 data preferred)\n   - Search: `[topic] study 2025 2026 data statistics`\n   - Prioritize tier 1-3 sources (see `skills\u002Fblog\u002Freferences\u002Fquality-scoring.md`)\n   - Record: statistic, source name, URL, date, methodology\n2. **Find a cover image** (wide, high-quality, topic-relevant):\n   - Search: `site:pixabay.com [topic] wide banner` (preferred)\n   - Alternative: `site:unsplash.com [topic] wide`\n   - Fallback: `site:pexels.com [topic] wide banner`\n   - Target dimensions: 1200x630 (OG-compatible) or 1920x1080\n   - Or generate a custom SVG cover via `blog-chart` (text-on-gradient with key stat)\n   - Or generate a custom AI image via `blog-image` sub-skill (if nanobanana-mcp configured)\n   - See `skills\u002Fblog\u002Freferences\u002Fvisual-media.md` for cover image sizing details\n3. **Find 3-5 inline images** from open-source platforms:\n   - **Pixabay** (preferred): Search `site:pixabay.com [topic keywords]`\n     - Extract image URL from page\n     - Direct URLs: `https:\u002F\u002Fcdn.pixabay.com\u002Fphoto\u002FYYYY\u002FMM\u002FDD\u002FHH\u002FMM\u002Ffilename.jpg`\n     - Verify with `curl -sI \"\u003Curl>\" | head -1` returns HTTP 200\n   - **Unsplash** (alternative): Search `site:unsplash.com [topic keywords]`\n     - Build URL: `https:\u002F\u002Fimages.unsplash.com\u002Fphoto-\u003Cid>?w=1200&h=630&fit=crop&q=80`\n   - **Pexels** (fallback): Search `site:pexels.com [topic keywords]`\n4. **Plan 2-4 data visualizations** from researched statistics\n   - Select diverse chart types (see `skills\u002Fblog\u002Freferences\u002Fvisual-media.md`)\n   - Map data points to chart formats\n5. **AI image generation** (optional, if nanobanana-mcp configured):\n   - If stock photo results are insufficient (\u003C 3 good matches) or topic is too niche\n   - Generate custom hero image and\u002For inline illustrations via `blog-image` sub-skill\n   - Stock photos remain default - AI generation supplements, never replaces\n6. **NotebookLM research** (optional, if user has relevant notebooks):\n   - If the user mentions a NotebookLM notebook or the topic aligns with a configured notebook\n   - Query via `blog-notebooklm` for source-grounded data from user-uploaded documents\n   - Treat NotebookLM responses as Tier 1 sources (user's own primary documents)\n   - Falls back silently if not configured or not authenticated\n7. **Find relevant YouTube videos** (2-3 per post):\n   - Use `blog-google` youtube command or WebSearch `site:youtube.com [topic] [year]`\n   - Apply quality criteria from `skills\u002Fblog\u002Freferences\u002Fvideo-embeds.md` (min score 50\u002F100)\n   - Select 2-3 best videos. Falls back silently if none found.\n\n### Phase 3: Outline Generation\n\nCreate a structured outline before writing. If a template was loaded in Phase 1.5,\nadapt this skeleton to match the template's section structure:\n\n```\n# [Title as Question - Include Primary Keyword]\n\n## Introduction (100-150 words)\n- Hook with surprising statistic\n- Problem\u002Fopportunity statement\n- What the reader will learn\n\n> **Key Takeaways**\n> - [Core finding with statistic and source]\n> - [Second key insight or recommendation]\n> - [Third actionable takeaway]\n> (3-5 bullets, 40-60 words combined)\n\n## H2: [Question Format] (300-400 words)\n- Answer-first paragraph (40-60 words with stat + source)\n- Supporting evidence\n- [Image placement]\n- Practical advice\n- [CITATION CAPSULE placeholder]\n- [INTERNAL-LINK: anchor text → target description]\n\n## H2: [Question Format] (300-400 words)\n- Answer-first paragraph\n- [Chart: type + data description]\n- Analysis and implications\n- [CITATION CAPSULE placeholder]\n- [INTERNAL-LINK: anchor text → target description]\n\n## H2: [Statement for Variety] (300-400 words)\n- Answer-first paragraph\n- Real-world example or case study\n- [Image placement]\n- [CITATION CAPSULE placeholder]\n\n## H2: [Question Format] (300-400 words)\n- Answer-first paragraph\n- [Chart: type + data description]\n- Step-by-step guidance\n- [CITATION CAPSULE placeholder]\n- [INTERNAL-LINK: anchor text → target description]\n\n## H2: [Question Format] (200-300 words)\n- Answer-first paragraph\n- Forward-looking analysis\n\n## [CTA Section or Inline Placement]\n- See `skills\u002Fblog\u002Freferences\u002Fcta-placement.md` for placement rules by content type\n- Place CTA after value delivery, not at arbitrary positions\n- Single focused CTA per post (266% more conversions)\n- [CTA: contextual call-to-action matching article topic]\n\n## FAQ Section (3-5 questions, 40-60 words each answer)\n- [INTERNAL-LINK: anchor text → detailed content]\n\n## Conclusion (100-150 words)\n- Key takeaways (bulleted)\n- Call to action\n- [INTERNAL-LINK: anchor text → next logical content]\n```\n\nPresent the outline to the user for approval before writing.\n\n**Visual element pacing**: Insert `[IMAGE]`, `[CHART]`, `[VIDEO]`, or `[CALLOUT]` markers\nevery 300-500 words. Alternate types (no consecutive same-type). See\n`skills\u002Fblog\u002Freferences\u002Fcontent-rules.md` Visual Rhythm section and\n`skills\u002Fblog\u002Freferences\u002Fcta-placement.md` for CTA positioning.\n\n### Phase 4: Chart Generation (Built-In)\n\nWhen the researcher identifies chart-worthy data (3+ comparable metrics, trend data,\nbefore\u002Fafter comparisons):\n\n1. Select chart type using the diversity rule (no repeated types per post)\n2. Invoke `blog-chart` sub-skill with: chart type, title, data values, source, platform format\n3. Embed the returned SVG directly in the post within a `\u003Cfigure>` wrapper\n4. Target 2-4 charts per 2,000-word post\n5. Distribute charts evenly - never cluster them\n\nSee `skills\u002Fblog\u002Freferences\u002Fvisual-media.md` for chart type selection and styling rules.\n\n### Phase 5: Content Writing\n\nWrite the full article following these rules:\n\n#### 5a. Frontmatter\n```yaml\n---\ntitle: \"[Question-format title with primary keyword]\"\ndescription: \"[Fact-dense, 150-160 chars, includes 1 statistic]\"\ncoverImage: \"[URL from Pixabay\u002FUnsplash\u002FPexels or generated SVG path]\"\ncoverImageAlt: \"[Descriptive sentence about the cover image]\"\nogImage: \"[Same as coverImage, or custom OG image URL]\"\ndate: \"YYYY-MM-DD\"\nlastUpdated: \"YYYY-MM-DD\"\nauthor: \"[Author name]\"\ntags: [\"keyword1\", \"keyword2\", \"keyword3\"]\n---\n```\n\nIf the platform uses a different field name (e.g., `image`, `hero`, `thumbnail`),\nadapt to match the project's existing frontmatter convention.\n\n#### 5b. Summary Box (Key Takeaways)\n\nImmediately after the introduction (before the first H2 body section), add a summary box:\n\n```markdown\n> **Key Takeaways**\n> - [Core finding with statistic] ([Source], year)\n> - [Second key insight or recommendation]\n> - [Third actionable takeaway]\n```\n\nRequirements:\n- 3-5 bullet points, 40-60 words combined\n- Must be self-contained - understandable without reading the article\n- Include 1 specific statistic with source name\n- State the key finding, recommendation, or answer\n- Default label: \"Key Takeaways\". If a persona is active, use the persona's summary_label\n- Backward compatible: accept existing TL;DR boxes during rewrites\n\n#### 5c. Answer-First Formatting (Critical)\nEvery H2 section MUST open with a 40-60 word paragraph containing:\n- At least one specific statistic with source attribution\n- A direct answer to the heading's implicit question\n\nPattern:\n```markdown\n## How Does X Impact Y in 2026?\n\n[Stat from source] ([Source Name](url), year). [Direct answer to the heading\nquestion in 1-2 more sentences, explaining the implication and what this means\nfor the reader.]\n```\n\n**FLOW evidence triple (drafting requirement, not just audit):**\n\nEvery public statistic must carry three components AT DRAFTING TIME:\n\n1. **Year anchor in prose.** Write \"In 2026,\" or \"As of Q1 2026,\" BEFORE\n   the statistic, in the sentence body. Year buried inside parentheses\n   does not count. Example:\n   - GOOD: \"In 2026, Ahrefs found a 58% lower CTR for position one when\n     an AI Overview was present.\"\n   - WEAK: \"Position-one CTR dropped 58% (Ahrefs, 2026).\"\n\n2. **Inline citation with publisher and title.** Name both the publisher\n   and the document title (or report name), not just a brand. Example:\n   - GOOD: \"Ahrefs, AI Overviews CTR update, December 2025\"\n   - WEAK: \"Ahrefs reported...\"\n\n3. **URL plus retrieval date in the source block at the bottom of the post.**\n   Provenance discipline lets future readers and AI crawlers verify the\n   source still says what was claimed. Format:\n   - \"[Publisher], [Title], retrieved YYYY-MM-DD, [full URL]\"\n\n**FLOW quality bar (drop or replace):**\nPublic claims must use verified sources OR stay qualitative. If a statistic\ncannot be verified, drop it. If it is contradicted by a more recent source,\nreplace it with the verified alternative. Do not soften vague language to\nkeep an unsourceable number.\n\nFor evidence-led optimization prompts (CTR audit, AI detector test, schema,\nPAA rewording, ChatGPT visibility), see `\u002Fblog flow optimize`.\n\n#### 5d. Information Gain Markers\n\nDistribute at least 2-3 information gain markers throughout the article. These\nsignal to search engines and AI systems that the content contains original value\nnot available elsewhere.\n\nTag each with a comment or visible marker:\n\n- `[ORIGINAL DATA]` - Proprietary surveys, experiments, A\u002FB test results, case\n  study metrics the author collected first-hand\n- `[PERSONAL EXPERIENCE]` - First-hand observations, lessons learned from direct\n  involvement, \"when we tried X, Y happened\" narratives\n- `[UNIQUE INSIGHT]` - Analysis others haven't made, contrarian perspectives\n  backed by data, novel connections between existing research\n\nPlacement:\n- Weave into the body text naturally\n- Use as inline comments: `\u003C!-- [ORIGINAL DATA] -->` before the relevant paragraph\n- Or as visible callouts if the format supports it:\n  ```markdown\n  > **Our finding:** [original observation backed by specific data]\n  ```\n- Minimum 2 per post, target 3 for comprehensive articles\n\nThese markers map directly to the \"Originality\u002Funique value markers\" criterion\nin the Content Quality scoring category (see `skills\u002Fblog\u002Freferences\u002Fquality-scoring.md`).\n\n#### 5e. Citation Capsules\n\nFor each major H2 section, generate a citation capsule - a 40-60 word self-contained\npassage designed so AI systems can extract and quote it directly.\n\nRequirements per capsule:\n- 40-60 words, self-contained (makes sense in isolation)\n- Contains: one specific claim + one data point + source attribution\n- Written in a declarative, quotable style\n- Placed within the H2 section body (not as a separate block)\n\nExample:\n```markdown\nAccording to a 2026 Gartner study, 58% of enterprise buyers now consult AI\nassistants before contacting a vendor ([Gartner](https:\u002F\u002Fwww.gartner.com), 2026).\nThis shift means B2B content must answer specific questions concisely enough\nfor AI systems to extract and cite in their responses.\n```\n\nCapsules map to the \"AI Citation Readiness\" scoring category (15 points) in\n`skills\u002Fblog\u002Freferences\u002Fquality-scoring.md`.\n\n#### 5f. Internal Linking Zones\n\nMark internal linking opportunities throughout the article using placeholder\nnotation. The user (or a follow-up pass) will resolve these to actual URLs.\n\nZone placement:\n- **Introduction** - Link to related pillar content or topic hub\n- **Each H2 section** - Link to supporting articles, deeper dives, related tools\n- **FAQ section** - Link answers to detailed content that expands on the answer\n- **Conclusion** - Link to the next logical piece of content the reader should consume\n\nFormat:\n```markdown\n[INTERNAL-LINK: anchor text → target description]\n```\n\nExample:\n```markdown\nFor a deeper dive into keyword clustering, see our\n[INTERNAL-LINK: complete guide to keyword clustering → pillar page on keyword research methodology].\n```\n\nTarget 5-10 internal link zones per 2,000-word post. Use descriptive anchor text\n(never \"click here\" or \"read more\"). See `skills\u002Fblog\u002Freferences\u002Finternal-linking.md` for\nanchor text rules and linking strategy.\n\n#### 5g. Paragraph Rules\n- Every paragraph: 40-80 words (never exceed 150)\n- Every sentence: max 15-20 words\n- Start each paragraph with the most important information\n- Target Flesch Reading Ease: 60-70\n\n#### 5h. Heading Rules\n- One H1 (title only)\n- H2s for main sections (60-70% as questions)\n- H3s for subsections only - never skip levels\n- Include primary keyword naturally in 2-3 headings\n\n#### 5i. Image Embedding\n\nStandard markdown:\n```markdown\n![Descriptive alt text - topic keywords naturally](https:\u002F\u002Fcdn.pixabay.com\u002Fphoto\u002F...)\n```\n\nMDX with Next.js Image (if detected):\n```mdx\n![Descriptive alt text - topic keywords naturally](https:\u002F\u002Fcdn.pixabay.com\u002Fphoto\u002F...)\n```\n\n- Place images after H2 headings, before body text\n- Space evenly throughout the post (not clustered)\n- Alt text should be a full descriptive sentence\n\n#### 5j. Chart Embedding\n\nStandard markdown\u002FHTML:\n```html\n\u003Cfigure>\n  \u003Csvg viewBox=\"0 0 560 380\" ...>...\u003C\u002Fsvg>\n  \u003Cfigcaption>Source: [Source Name], [Year]\u003C\u002Ffigcaption>\n\u003C\u002Ffigure>\n```\n\nMDX format:\n```mdx\n\u003Cfigure className=\"chart-container\" style={{margin: '2.5rem 0', textAlign: 'center', padding: '1.5rem', borderRadius: '12px'}}>\n  \u003Csvg viewBox=\"0 0 560 380\" ...>...\u003C\u002Fsvg>\n\u003C\u002Ffigure>\n```\n\n#### 5k. Video Embedding\nEmbed YouTube videos using srcdoc lazy-loading pattern from `skills\u002Fblog\u002Freferences\u002Fvideo-embeds.md`.\nInclude aria-label, noscript fallback for AI crawlers. Place after relevant H2, 500+ words apart.\n\n#### 5l. Citation Format\nInline attribution (always):\n```markdown\nOrganic CTR declined 61% with AI Overviews ([Seer Interactive](https:\u002F\u002Fwww.seerinteractive.com\u002F), 2025).\n```\n\n#### 5m. FAQ Section\nAdd 3-5 FAQ items with 40-60 word answers. Each answer must contain a statistic.\n\nFor MDX with FAQSchema component:\n```mdx\n\u003CFAQSchema faqs={[\n  { question: \"Question?\", answer: \"40-60 word answer with statistic and source.\" },\n]} \u002F>\n```\n\nFor standard markdown:\n```markdown\n## Frequently Asked Questions\n\n### Question text here?\n\nAnswer with statistic and source attribution (40-60 words).\n```\n\n#### 5n. Internal Linking\n- 5-10 internal links per 2,000-word post\n- Link to relevant existing content naturally\n- Use descriptive anchor text (not \"click here\")\n\n### Phase 6: Quality Check\n\nBefore delivering, verify:\n\n#### Structure and Content\n1. Every H2 opens with a statistic + source\n2. No paragraph exceeds 150 words\n3. All statistics have named tier 1-3 sources\n4. 2-4 charts with type diversity\n5. 3-5 inline images with descriptive alt text\n6. Cover image present in frontmatter (coverImage + ogImage)\n7. FAQ section present with 3-5 items\n8. Heading hierarchy is clean (H1 -> H2 -> H3)\n9. Meta description is 150-160 chars with a stat\n\n#### New Element Verification\n10. TL;DR box present after introduction (40-60 words, contains statistic + source)\n11. At least 2-3 information gain markers (`[ORIGINAL DATA]`, `[PERSONAL EXPERIENCE]`, or `[UNIQUE INSIGHT]`)\n12. Citation capsules present in major H2 sections (40-60 words, self-contained, quotable)\n13. Internal linking zones marked in introduction, H2 sections, FAQ, and conclusion\n14. No AI-detectable phrases from banned list (see `agents\u002Fblog-writer.md`)\n\n#### Burstiness and Naturalness Check\n15. **Sentence length variance** - Verify a mix of short (8-word) and long (25-word) sentences. Uniform sentence length signals AI authorship.\n16. **Banned AI phrase scan** - Check for and remove:\n    - \"in today's digital landscape\", \"it's important to note\", \"dive into\"\n    - \"game-changer\", \"navigate the landscape\", \"revolutionize\", \"seamlessly\"\n    - \"cutting-edge\", \"harness the power of\", \"leverage\" (as verb)\n    - \"delve\", \"crucial\", \"elevate\", \"foster\", \"landscape\" (overused)\n    - \"multifaceted\", \"robust\", \"tapestry\", \"embark\"\n    - Full list in `agents\u002Fblog-writer.md`\n17. **Contractions** - Verify natural use of contractions (\"it's\", \"we've\", \"don't\", \"isn't\"). Formal AI prose avoids contractions; natural writing uses them.\n18. **Rhetorical questions** - Verify at least one rhetorical question every 200-300 words to break up declarative patterns.\n19. **YouTube videos** - 2-3 embeds with lazy loading, aria-labels, and noscript fallback (see `skills\u002Fblog\u002Freferences\u002Fvideo-embeds.md`)\n\n### Phase 6.5: Delivery Contract Enforcement (v1.9.0)\n\nBefore Phase 7, run the 5-gate delivery contract per `skills\u002Fblog\u002Freferences\u002Fblog-delivery-contract.md`. The user is never the first reviewer; the gates are.\n\nSteps:\n\n1. **Capability discovery + hero**: run `python scripts\u002Fblog_preflight.py --draft \u003Cfolder> --gate 1` to enumerate available paths. If `nanobanana-mcp` is loaded, generate the hero via the MCP tool. Otherwise run `python scripts\u002Fgenerate_hero.py --topic \"\u003Ctitle>\" --tags \"\u003Ctags>\" --out \u003Cfolder>` (uses the Gemini, Unsplash, Pexels, Pixabay, Openverse ladder).\n\n2. **Format completeness**: render the canonical `.md` to `.html` and `.pdf` via `python scripts\u002Fblog_render.py --md \u003Cslug>.md --out-dir \u003Cfolder>`. All three artifacts plus `hero.\u003Cext>` must end up in the draft folder.\n\n3. **Content review (blocking)**: dispatch the `blog-reviewer` agent (Task tool) against the rendered `.html`. The agent emits its scorecard to `\u003Cfolder>\u002Freview.md` ending with `BLOCKING: true|false (reason)`. Threshold: overall score 90\u002F100 or higher AND zero P0 issues per `editorial-heuristics.md`.\n\n4. **Visual + asset gates**: run `python scripts\u002Fblog_preflight.py --draft \u003Cfolder> --strict`. This runs Gate 3 (visual verification via patchright at 3 viewport widths), Gate 4 (reads review.md BLOCKING line), and Gate 5 (asset + link integrity). Exit 0 = ship; exit 1 = block.\n\n5. **Iteration**: on any block, capture the failure diagnostic from `\u003Cfolder>\u002Fpreflight-report.json`, re-dispatch the blog-writer agent with the diagnostic as input, and re-run from step 1. Maximum 3 iterations. On the 3rd failure, STOP and present the failure diagnostic instead of the draft.\n\nThe orchestrator holds the loop counter; this sub-skill never loops itself.\n\n### Phase 7: Delivery\n\nPresent the completed article ONLY after Phase 6.5 returns all gates passing. Include the screenshots from `\u003Cfolder>\u002Fpreview\u002F*.png` in the summary so the user can see what they are getting before reading the prose.\n\nSummary template:\n\n```\n## Blog Post Complete: [Title]\n\n### Template Used\n- [Template name] (or \"generic outline - no template matched\")\n\n### Statistics\n- [N] sourced statistics from tier 1-3 sources\n- [N] unique sources cited\n\n### Visual Elements\n- Cover image: [source - Pixabay\u002FUnsplash\u002FPexels or generated SVG]\n- [N] inline images (Pixabay\u002FUnsplash\u002FPexels)\n- [N] SVG charts (types: bar, lollipop, donut, line)\n- [N] YouTube video embeds (titles: ...)\n\n### Dual-Optimization Elements\n- TL;DR box: present (N words)\n- Information gain markers: [N] ([types used])\n- Citation capsules: [N] across H2 sections\n- Internal linking zones: [N] marked\n\n### Structure\n- [N] H2 sections with answer-first formatting\n- [N] FAQ items with schema\n- Word count: ~[N] words\n- Estimated reading time: [N] min\n\n### Naturalness\n- Sentence length variance: [pass\u002Ffail]\n- AI phrase scan: [pass\u002Ffail]\n- Contractions used: [yes\u002Fno]\n- Rhetorical questions: [N] (target: 1 per 200-300 words)\n\n### Next Steps\n- Review and customize for your brand voice\n- Resolve [INTERNAL-LINK] placeholders with actual URLs\n- Add internal links to your existing content\n- Run `\u002Fblog analyze \u003Cfile>` to verify quality score\n- Generate VideoObject schema: `\u002Fblog schema \u003Cfile>` (includes video markup)\n- Generate audio narration: `\u002Fblog audio generate \u003Cfile>` (optional)\n```\n",{"name":5,"license":15,"description":24,"argument-hint":25,"user-invokable":26},"Write new blog articles from scratch optimized for Google rankings and AI citations. Generates full articles with template selection, answer-first formatting, Key Takeaways summary box, information gain markers, citation capsules, sourced statistics, Pixabay\u002FUnsplash images, built-in SVG chart generation, FAQ schema, internal linking zones, and proper heading hierarchy. Supports MDX, markdown, and HTML output. Use when user says \"write blog\", \"new blog post\", \"create article\", \"write about\", \"draft blog\", \"generate blog post\".\n","\u003Ctopic>",true,"main","8859218955d9504cd18b59e7892100718039d3b4","data\u002Fskills\u002Fagricidaniel-claude-blog\u002Fblog-write\u002Fskills\u002Fblog-write",1,"https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper","https:\u002F\u002Fgithub.com\u002Fom-scogo","SDE-1 @ Scogo ","2026-07-18 12:22:46",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":26},[37,40,43],{"tool":38,"command":39},"npx","npx skills add om-scogo\u002Fskillsh-scraper --skill blog-write",{"tool":41,"command":42},"bunx","bunx skills add om-scogo\u002Fskillsh-scraper --skill blog-write",{"tool":44,"command":45},"pnpm","pnpm dlx skills add om-scogo\u002Fskillsh-scraper --skill blog-write","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fagricidaniel-claude-blog\u002Fblog-write\u002Fskills\u002Fblog-write\u002FSKILL.md",[],[49,56,61,68,73,80],{"skillNo":50,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":19},"01M7D6MC9FAAEV4CMNYX8W7JTG","arxiv-search","Search arXiv physics, math, and computer science preprints using natural language queries. Powered by Valyu semantic search.","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fyorkeccak-scientific-skills\u002Farxiv-search\u002Fskills\u002Farxiv-search\u002FSKILL.md",[],[],{"skillNo":57,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":58,"categories":59,"keywords":60,"publishedAt":19},"01M7D6M85CKKN87TA98VCQV4H7","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fyorkeccak-scientific-skills\u002Farxiv-search\u002FSkill.md",[],[],{"skillNo":62,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":19},"01M7D4ZTCS4DR3CE4JY06X84T2","working-with-legacy-code","Safely change and test untested codebases using Feathers' \"Working Effectively with Legacy Code\". Use when the user mentions \"legacy code\", \"no tests\", \"untested codebase\", \"how do I test this\", \"seams\", \"characterization tests\", \"golden master\", \"sprout method\", \"afraid to change this code\", \"monster method\", \"dependency breaking\", or \"inherited a messy codebase\". Also trigger when changing code without tests safely, getting a class under test when constructors, statics, or singletons block it, adding features to tangled modules, or planning incremental test coverage for an old codebase. Covers the legacy-code change algorithm, seams, characterization tests, sprout\u002Fwrap, and dependency-breaking techniques. For refactoring code that already has tests, see refactoring-patterns. For day-to-day code quality, see clean-code.","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fwondelai-skills\u002Fworking-with-legacy-code\u002Fworking-with-legacy-code\u002FSKILL.md",[],[],{"skillNo":69,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M7D4ZHZKJD2RSNGAF63V14P4","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fwondelai-skills\u002Fworking-with-legacy-code\u002FSkill.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M7D4ZB4CJBFW5JE67CPH2QRV","web-typography","Select, pair, and implement typefaces for web projects. Use when the user mentions \"font pairing\", \"which typeface\", \"line height\", \"responsive typography\", \"web font loading\", \"type hierarchy\", \"variable fonts\", \"FOUT\u002FFOIT\", \"typographic scale\", or \"the text is hard to read\". Also trigger when choosing between system fonts and web fonts, optimizing font-loading performance, or designing readable long-form content. Covers readability evaluation, CSS implementation, and performance optimization. For overall UI design systems, see refactoring-ui. For dramatic typographic experiences, see top-design.","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fwondelai-skills\u002Fweb-typography\u002Fweb-typography\u002FSKILL.md",[],[],{"skillNo":81,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7D4Z4MVYVY3NMJ18BNV6WNA","https:\u002F\u002Fgithub.com\u002Fom-scogo\u002Fskillsh-scraper\u002Fblob\u002Fmain\u002Fdata\u002Fskills\u002Fwondelai-skills\u002Fweb-typography\u002FSkill.md",[],[]]