Programmatic SEO page factory with thin-content guardrails
Problem solved: Programmatic SEO defaults to thin city-swap pages and gets penalized; a 12-playbook selector plus data-source mapping and CMS provisioning enforces quality over quantity.
What
A skill that designs and provisions programmatic-SEO page systems at scale: it chooses one of 12 page playbooks (Templates / Curation / Conversions / Comparisons / Examples / Locations / Personas / Integrations / Glossary / Translations / Directory / Profiles), creates the matching CMS collection in Webflow via MCP, populates sample or imported data, and produces a binding guide for the Webflow Designer template page — with a pre-launch quality checklist and post-launch monitoring built in.
Why it matters
Programmatic SEO is usually a vibes-based exercise that generates thousands of thin city-swap pages and earns a manual penalty. The 12-playbook taxonomy plus a data-source → playbook selector turns it into a structured decision keyed on the explicit principle "Better 100 great pages than 10,000 thin ones" — codifying the difference between a defensible page system and one that gets de-indexed.
End-to-end
- Discovery. Check for
.claude/product-marketing-context.md; confirm business context, opportunity assessment (search patterns, page-count potential), and competitive landscape (can you realistically compete). - Strategy & playbook selection. Apply the core principles (unique value per page; proprietary data > UGC > licensed > public; subfolders not subdomains; search-intent match; quality over quantity) and map data assets to a playbook via the selector table (proprietary data → Directory/Profiles; integrations → Integrations; local presence → Locations; competitor landscape → Comparisons).
- Webflow CMS setup. Use Webflow MCP tools (
data_sites_tool > list_sites,data_cms_tool) to create the playbook-specific collection with its recommended schema (Locations needs City/State/Hero Image/Local Description/Meta Title+Desc; Comparisons needs Product A/B Names/Winner/Feature Comparison/Meta). - Data population. Offer three options: generate 3–5 realistic sample entries, bulk import (CSV/JSON/structured), or user-adds-manually-later.
- Template page guidance. CMS template pages can't be created via API — provide step-by-step Webflow Designer instructions: bind Hero H1 to Name/Title, Rich Text to main content, SEO Title to Meta Title, OG Image to Hero Image; add a filtered Collection List at the bottom for internal linking (same category, limit 3–6, exclude current item).
- Pre-launch checklist. Unique value per page, intent answered, unique titles/meta, proper H1, schema markup, page speed, connected to site architecture, no orphans, in XML sitemap, no conflicting noindex.
- Post-launch monitoring. Indexation rate, rankings, traffic, engagement, conversion; watch for thin-content warnings, ranking drops, manual actions, crawl errors.
Prompts
Core Principles
1. Unique Value Per Page - Every page must provide value specific to that page
2. Proprietary Data Wins - First-party data > User-generated > Licensed > Public
3. Clean URL Structure - Always use subfolders, not subdomains
4. Search Intent Match - Pages must answer what people are searching for
5. Quality Over Quantity - Better 100 great pages than 10,000 thin ones
Gotchas
CMS template pages cannot be created via the Webflow API — that step is unavoidably manual in the Designer, so the workflow is not fully headless and budget time for it. The named anti-patterns are concrete: thin content (city-swap only), keyword cannibalization, over-generation without demand, poor data quality. Generating before you have a real proprietary data source is the fast path to a penalty.
Tools
- Claude Code with the skill installed; Webflow MCP server connected (
data_sites_tool,data_cms_tool); a Webflow account + site - Install:
/plugin marketplace add naveedharri/benai-skills && /plugin install marketing@benai-skills - Note: repo default branch is
develop, notmain— matters if cloning manually