content-seo
Copy + discoverability requirements every delivered app must meet — real-brand extraction, anti-slop voice, per-page SEO/JSON-LD, pSEO, GEO/AI-search, truth/citations, competitor floor, i18n, trust surfaces.
Install
npx skills add https://github.com/heymegabyte/agent-skills/tree/main/23-content-seo
claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install heymegabyte-agent-skills@llmmart
git clone https://github.com/heymegabyte/agent-skills.git
The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole heymegabyte/agent-skills collection as a plugin from our marketplace. Git is the plain clone.
Skill manifest
Content & SEO — every app's copy + discoverability must satisfy
Brand extraction (rebuilds / clients)
- Extract the REAL brand (logo, colors, fonts, tone; Wayback if the source is down). NEVER invent colors, never use Emdash defaults for clients, never infer from category.
- Colors derive from the logo (hue-distance verified). Logo luminance + source-site theme decide light/dark. Every logo render contrasts its container ≥4.5:1. Exactly ONE logo asset per container.
Anti-AI-slop (mandatory scan)
- Banned words: delve, leverage, unleash, revolutionize, seamless, robust, synergy, elevate, empower, transformative, best-in-class, cutting-edge, discover, innovative.
- Banned patterns: "Welcome to", "Discover [product]", vague aspirational headlines, hedging ("may help"), generic superlatives. Test: "Would the founder actually say this?" — no → rewrite.
Copy
- Headlines benefit-first, specific, ≤8 words. Subheads 15-25 words. Body one idea/paragraph, active, concrete. Max sentence 25 words. Flesch ≥60 (microcopy ≥70).
- CTAs: action-verb-first, ≤3 words preferred; never "Click here" / "Submit" / "Learn more". Above fold + page end.
- Microcopy: errors = [what happened] + [what to do]; empty states suggest the first action; toasts past-tense success / present in-progress.
SEO — every page
- Title 50-60 chars keyphrase-first · meta description 120-156 · canonical · OG image 1200×630 · exactly one H1 in prerendered HTML.
- JSON-LD: WebPage floor; Org/BreadcrumbList/FAQPage/Person/Product/Service ONLY when the entity is real — never pad (FAQPage only with real FAQs).
- sitemap.xml (+lastmod) · robots.txt · llms.txt · humans.txt · security.txt.
Blog & Reference Editorial Publishing (@theme-tools/blog — skills/blog-publishing/SKILL.md)
- 2,000+ Substantive Editorial Words (
D01–D03, NON-NEGOTIABLE): Every newly authored or generated blog / reference article MUST contain at least 2,000 substantive editorial words (target range 2,000–3,000+ words; navigation, bibliography, table cells, captions, code blocks, alt text, and boilerplate are excluded fromcomputeSubstantiveWordCount). Never settle for 600–900 word thin posts. Historical imported articles preserve or expand their original length (H28). - 8-Stage Editorial Pipeline & 12-Point Release Gate (
runQualityAndReleaseGate): Topic Discovery (15-idea topic strategy) → Live DataForSEO v3 Brief (isFixture: false) → SSRF-Safe Evidence & Source Ledger (SourceVersion+EvidenceReference) → Argument Architecture & Outline → 2,000–3,000+ Word Plate/Slate Draft (ValidatedPlateDocument) → ~12 Substantive Visual Placements (E01–E30: custom images, deterministic SVG charts, comparison tables, interactive calculators, product captures) → Inline Claim-Level Citations (CitationNode[n]+ bottomBibliographyWidget) → 12-Point Quality & Release Gate (runQualityAndReleaseGate) + 3 Daily Publishing Slots (09:00,14:00,19:00America/New_York). - Always load and follow
~/.agentskills/skills/blog-publishing/SKILL.md(andreferences/editorial-writing.md,references/visual-production.md,references/dataforseo.md,requirement-manifest.json) whenever planning, writing, auditing, or building/blogand/blog/:slug.
pSEO (when used)
- 5 page types: integration · comparison · use-case · template · location. Each: unique H1 + meta + 1,500–2,000+ unique words + ≥1 unique image + ≥3 internal links + ≥1 outbound citation. Never templatize verbatim.
GEO / AI search
- Quotable answer blocks (40-60 words). FAQPage schema. Every JSON-LD fact also appears as visible HTML. Lead paragraph answers the query in <40 words. EEAT: author bio + Person schema +
sameAs+ dated revisions + ownership.
Truth & citations
- Every factual claim APA-cited to a real source. Never fabricate people, stats, timelines, or testimonials. Testimonials carry verifiable attribution.
Reach
- Competitor floor: outscore the top 5-10 peer sites on every rubric dimension by ≥15% before "done".
- i18n by demographics: any community ≥10% local share (ACS B16001) gets a
/{locale}/*mirror + hreflang.
Required pages
/privacy·/terms·/accessibility· About/Team with real bios.
Files (agent-skills)
-
SKILL.md 4.6 KB
--- name: "content-seo" description: "Copy + discoverability requirements every delivered app must meet — real-brand extraction, anti-slop voice, per-page SEO/JSON-LD, pSEO, GEO/AI-search, truth/citations, competitor floor, i18n, trust surfaces." triggers: - "content" - "copy" - "seo" - "brand" - "blog" priority: 2 pack: "content" stage: stable --- # Content & SEO — every app's copy + discoverability must satisfy ## Brand extraction (rebuilds / clients) - Extract the REAL brand (logo, colors, fonts, tone; Wayback if the source is down). NEVER invent colors, never use Emdash defaults for clients, never infer from category. - Colors derive from the logo (hue-distance verified). Logo luminance + source-site theme decide light/dark. Every logo render contrasts its container ≥4.5:1. Exactly ONE logo asset per container. ## Anti-AI-slop (mandatory scan) - Banned words: delve, leverage, unleash, revolutionize, seamless, robust, synergy, elevate, empower, transformative, best-in-class, cutting-edge, discover, innovative. - Banned patterns: "Welcome to", "Discover [product]", vague aspirational headlines, hedging ("may help"), generic superlatives. Test: "Would the founder actually say this?" — no → rewrite. ## Copy - Headlines benefit-first, specific, ≤8 words. Subheads 15-25 words. Body one idea/paragraph, active, concrete. Max sentence 25 words. Flesch ≥60 (microcopy ≥70). - CTAs: action-verb-first, ≤3 words preferred; never "Click here" / "Submit" / "Learn more". Above fold + page end. - Microcopy: errors = [what happened] + [what to do]; empty states suggest the first action; toasts past-tense success / present in-progress. ## SEO — every page - Title 50-60 chars keyphrase-first · meta description 120-156 · canonical · OG image 1200×630 · exactly one H1 in prerendered HTML. - JSON-LD: WebPage floor; Org/BreadcrumbList/FAQPage/Person/Product/Service ONLY when the entity is real — never pad (FAQPage only with real FAQs). - sitemap.xml (+lastmod) · robots.txt · llms.txt · humans.txt · security.txt. ## Blog & Reference Editorial Publishing (`@theme-tools/blog` — `skills/blog-publishing/SKILL.md`) - **2,000+ Substantive Editorial Words (`D01–D03`, NON-NEGOTIABLE)**: Every newly authored or generated blog / reference article MUST contain **at least 2,000 substantive editorial words** (target range **2,000–3,000+ words**; navigation, bibliography, table cells, captions, code blocks, alt text, and boilerplate are excluded from `computeSubstantiveWordCount`). **Never settle for 600–900 word thin posts.** Historical imported articles preserve or expand their original length (`H28`). - **8-Stage Editorial Pipeline & 12-Point Release Gate (`runQualityAndReleaseGate`)**: Topic Discovery (`15-idea topic strategy`) → Live DataForSEO v3 Brief (`isFixture: false`) → SSRF-Safe Evidence & Source Ledger (`SourceVersion` + `EvidenceReference`) → Argument Architecture & Outline → 2,000–3,000+ Word Plate/Slate Draft (`ValidatedPlateDocument`) → ~12 Substantive Visual Placements (`E01–E30`: custom images, deterministic SVG charts, comparison tables, interactive calculators, product captures) → Inline Claim-Level Citations (`CitationNode` `[n]` + bottom `BibliographyWidget`) → 12-Point Quality & Release Gate (`runQualityAndReleaseGate`) + 3 Daily Publishing Slots (`09:00`, `14:00`, `19:00` `America/New_York`). - Always load and follow `~/.agentskills/skills/blog-publishing/SKILL.md` (and `references/editorial-writing.md`, `references/visual-production.md`, `references/dataforseo.md`, `requirement-manifest.json`) whenever planning, writing, auditing, or building `/blog` and `/blog/:slug`. ## pSEO (when used) - 5 page types: integration · comparison · use-case · template · location. Each: unique H1 + meta + 1,500–2,000+ unique words + ≥1 unique image + ≥3 internal links + ≥1 outbound citation. Never templatize verbatim. ## GEO / AI search - Quotable answer blocks (40-60 words). FAQPage schema. Every JSON-LD fact also appears as visible HTML. Lead paragraph answers the query in <40 words. EEAT: author bio + Person schema + `sameAs` + dated revisions + ownership. ## Truth & citations - Every factual claim APA-cited to a real source. Never fabricate people, stats, timelines, or testimonials. Testimonials carry verifiable attribution. ## Reach - Competitor floor: outscore the top 5-10 peer sites on every rubric dimension by ≥15% before "done". - i18n by demographics: any community ≥10% local share (ACS B16001) gets a `/{locale}/*` mirror + hreflang. ## Required pages - `/privacy` · `/terms` · `/accessibility` · About/Team with real bios.
Comments (0)
Sign in to join the conversation.
Reviews (0)
No reviews yet.
No comments yet.