Marketer pack
Claude Skill
Programmatic SEO Builder
Plans SEO pages at scale from a template plus data — location, comparison, integration pages — without tripping thin-content penalties.
What it does
Designs a programmatic SEO system: identifies the repeatable search pattern ([keyword]+[city], comparisons, integration pages), assesses whether you can realistically rank for it, and specifies a template that delivers genuine unique value per page so you can scale to hundreds of pages without thin-content penalties.
When to use
- ✓There's a search pattern with many variations (locations, integrations, comparisons)
- ✓You have, or can get, the data to fill a template with real value
- ✓You want SEO traffic at scale, not page by page
When not to use
- ✗You can't make each page genuinely useful — you'll get penalized, not ranked
- ✗It's a handful of pages — just write them directly
Install
Download the .zip, then unzip into your Claude skills folder.
mkdir -p ~/.claude/skills
unzip ~/Downloads/programmatic-seo-builder.zip -d ~/.claude/skills/
# Restart Claude Code session.
# Skill is now available — Claude will use it when relevant.SKILL.md
SKILL.md
---
name: programmatic-seo-builder
description: Use when creating SEO-driven pages at scale from templates and data. Triggers on "programmatic SEO", "pSEO", "pages at scale", "location pages", "[keyword]+[city] pages", "integration pages", "generate 100 pages".
---
# Programmatic SEO Builder
Programmatic SEO is building many similar pages from a template and a dataset. The line between a ranking asset and a spam penalty is one thing: unique value per page. Get that right and you scale; get it wrong and Google buries the whole set as thin content.
## Unique value per page
Each page must answer the specific query better than a generic page would — with real, page-specific data, not the same paragraph with a city name swapped in. If a human couldn't tell two of your pages apart except for one variable, it's thin. Spun text is the fastest way to get the set deindexed.
## Opportunity assessment first
Before building the template, check the pattern is worth it:
- **Search pattern** — is there genuine, repeated search demand across the variations?
- **Volume distribution** — a few high-volume + a long tail, or all near-zero?
- **Can you compete** — who ranks now, and can your pages realistically beat them?
No demand or no chance to compete means no pages.
## The template
Separate what's templated (structure, layout) from what must be unique per page (the actual data, stats, examples, local detail). The unique part is the product. Source real data for it; don't generate filler.
## Rollout
Don't dump 500 pages at once — publish in batches, confirm they index and hold quality, then scale. Internal-link the set so the pages reinforce each other and route to priority pages.
## Anti-patterns
- Thin templated pages distinguishable only by one variable
- Spun or AI-filler text instead of real per-page data
- Mass-publishing the whole set before checking quality and indexing
- Building before assessing whether you can realistically rank
Example prompts
Once installed, try these prompts in Claude:
- We integrate with 80 tools. Plan programmatic /integrations/[tool] pages that can actually rank.
- Plan location pages for a service business across 200 cities without thin content.