SkillAgentSearch skills...

programmatic-seo

How to design and run a programmatic SEO program that produces durable traffic instead of penalty-bait. Data source identification, template design, schema patterns, quality control at scale, internal linking architecture, crawl budget management, AEO/GEO for programmatic pages, refresh discipline,…

Install / Use

npx skills add rampstackco/claude-skills --skill programmatic-seo

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

86/100

Category

Marketing

Supported Platforms

Universal

Our assessment of programmatic-seo

programmatic-seo scores 86/100 on our quality scale, 413th of 642 Marketing skills we index.

Its SKILL.md is 24 KB long, well organised into 16 sections and no code examples: a thorough specification that gives an agent plenty to work with.

It has 902 GitHub stars, a meaningful sign that others use it.

Substance
30/30
Structure
13/20
Description
15/15
Adoption
13/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 2 months ago, so programmatic-seo is actively maintained.
  • Our last check on 2026-11-03 found the source still online.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-07. Automated pattern scan on 2026-10-07. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

programmatic-seo compared with similar skills

All 4 of these similar skills score higher than programmatic-seo; compare them before choosing.

SkillScoreStarsUpdatedFormat
programmatic-seo (this skill)by rampstackco869022mo agoSKILL.md
taste-skillby Leonxlnx100116.8k8d agoCLAUDE.md
Claude-Code-Game-Studiosby Donchitos10039.7k7d agoCLAUDE.md
x402by browser-use100121.0k3d agoSKILL.md
algorithmic-artby anthropics100177.9k55d agoSKILL.md

Frequently asked questions

How do I install programmatic-seo?
Run npx skills add rampstackco/claude-skills --skill programmatic-seo. The install tabs above show the steps for each supported agent.
Which AI agents does programmatic-seo work with?
It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
Is programmatic-seo safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is programmatic-seo still maintained?
The repository was last updated about 2 months ago, so programmatic-seo is actively maintained.

name: programmatic-seo description: "How to design and run a programmatic SEO program that produces durable traffic instead of penalty-bait. Data source identification, template design, schema patterns, quality control at scale, internal linking architecture, crawl budget management, AEO/GEO for programmatic pages, refresh discipline, and the make-or-break question of whether pSEO is the right answer for your program at all. Triggers on programmatic SEO, pSEO, scaled content, page generation, template SEO, location pages, comparison pages, directory site, listing site, scaled landing pages, programmatic content. Also triggers when a content set is not ranking, has been hit by an algorithm update, or when a team is considering pSEO as a growth lever." category: content catalog_summary: "Designing pSEO programs that work: data sources, template design, quality control at scale, internal linking, crawl budget, AEO/GEO patterns, refresh discipline, and when pSEO is and is not the right answer" display_order: 6

Programmatic SEO

A senior SEO strategist's playbook for designing and running programmatic SEO programs that produce durable traffic instead of penalty-bait.

Programmatic SEO has a complicated reputation. Sites like Zillow, Airbnb, TripAdvisor, Indeed, and Yelp have built billion-dollar traffic engines on pSEO. Other sites have built pSEO programs that got hit by Google's helpful-content updates and lost 80% of their traffic in a week. The difference is rarely the technique; it is the underlying data quality and the quality control discipline at scale.

This skill is the playbook for getting that distinction right. It assumes you have decided what keyword space to target (see seo-keyword) and how the broader content program is shaped (see content-strategy). It does not write individual editorial pieces (see content-and-copy for that) and does not architect editorial topic hubs (see pillar-content-architecture for that). What it does is teach the discipline of generating high-volume pages programmatically from structured data sources without producing thin content, duplicate content, or scale-without-substance pages that get penalized.

When to use this skill: deciding whether pSEO is a fit for the program at all (the most important question), designing a new pSEO system, auditing an existing pSEO set that is not ranking or has been hit by an algorithm update, or building quality-control discipline for a pSEO program that grew faster than its quality processes.


What this skill is for

This skill spans scaled-content programs from data source through quality control. It composes with five sister and adjacent skills, and the distinction between them is what keeps each one sharp.

  • content-strategy is program-scope: editorial pillars, calendar, governance. Decides whether pSEO fits the program at all.
  • pillar-content-architecture is hub-scope: one topic with 10 to 15 intentional editorial pieces. Editorial in nature.
  • content-brief-authoring is per-piece scope: brief for one editorial artifact.
  • content-and-copy is execution scope: writing individual editorial pieces.
  • This skill is scaled scope: 100s to 100,000s of pages generated programmatically from structured data sources, each targeting a long-tail query.

The clean reading order: content-strategy decides whether pSEO is a fit, seo-keyword surfaces the long-tail keyword space, this skill designs the pSEO system, editorial-qa (forthcoming) provides the QA discipline for sampled quality control across the set. content-and-copy and content-brief-authoring are not in the loop for pSEO at scale; those skills are for editorial pieces, not data-driven generated pages.

The audience: SEO content strategists, content engineers, agencies running pSEO programs, in-house teams considering pSEO as a growth lever. The voice is senior SEO strategist to junior PM or marketer. Specific, opinionated, honest. The reputation problem is not pSEO; it is pSEO without underlying value.


When pSEO is the right answer

The keystone question. pSEO works when all of the following are true.

1. Real underlying data. A genuine structured data source with depth: 10+ fields per record, ideally 20+, with first-party data, expert-curated data, or licensed datasets. Not just scraped data or AI-generated facts dressed as structured records.

2. Long-tail query volume justifies the effort. The queries the program would target have meaningful aggregate volume even if individual queries are small. Real estate "homes for sale in {neighborhood}" works at scale. "Blue widget reviews 2026" generated for every adjective times widget combination does not, because the queries are not actually searched.

3. User intent is queryable. The user's question can be answered through structured data presented well. Not through narrative explanation, judgment, or analysis the data cannot supply.

4. Update cadence aligns with query volatility. Real estate listings update daily and the data refresh aligns. "Best [thing] for [year]" pages get stale annually and need a refresh discipline budgeted in.

5. Quality control is operationally feasible. The team has the capacity to sample-audit the set, fix failures, and maintain quality as the set grows. Not aspirationally; budgeted in headcount.

pSEO does NOT work when:

  • The underlying data is shallow (3 to 5 fields, mostly AI-generated, no first-party signal)
  • The query volume is illusory (long-tail keywords nobody actually searches)
  • User intent requires narrative or judgment that data cannot supply
  • Quality control is not budgeted (write a bunch of pages, ignore them)
  • The program is scaling AI-generated thin pages without unique data

The honest framing. Most teams that ask "should we do pSEO?" should hear "probably not, unless the underlying data is unique or first-party expertise makes the pages actually useful." The reputation problem is not the technique; it is the technique applied without underlying value.

Detail in references/when-pseo-works-decision.md.


Data source identification

The data source is the pSEO program. Common sources with different defensibility profiles.

First-party data. Customer transactions, content database, user-generated content. Defensible because nobody else has it. Examples: Glassdoor's employee reviews, Yelp's user ratings, TripAdvisor's traveler reviews.

Licensed datasets. Industry databases, regulatory data, government datasets, licensed third-party feeds. Defensible by license terms and integration depth. Examples: Zillow's MLS partnerships, real estate brokerage feeds, sports statistics licenses.

Aggregated public data. Scraped, cleaned, enriched. Judgment call on legality (often gray area depending on robots.txt, terms of service, jurisdictional rules). Defensibility depends on the cleaning and enrichment work. Easy to copy if the cleaning is shallow.

Expert-curated content. The dataset is built by hiring experts to populate it. Slow, high-quality, defensible. Examples: Wirecutter's product testing, expert-reviewed medical content, curated editorial databases.

Synthesized data. Combining multiple sources into a unique view. "Neighborhoods times schools times prices" combining three datasets into a comparison view. Defensibility comes from the synthesis logic and the ongoing maintenance of multi-source pipelines.

The "moat" question. Would a competitor be able to replicate the data source? If yes, the pSEO program has no defensibility; anyone can copy. If no, the data source becomes a moat that compounds. Zillow's MLS partnerships, Glassdoor's employee reviews, Crunchbase's funding data, are moats. "Scraped Wikipedia plus AI rewrite" is not.

Detail in references/data-source-identification-patterns.md.


Template design

The template is the structure that data fills. Design principles.

Above-the-fold answer. The user's specific question answered in the first 200 words of the rendered page, structured for both human reading and AI extraction. The answer is what gets cited; the rest is supporting depth.

Variable density. The template accommodates records with sparse fields (some have 5 data points, others have 50) without looking broken. Sparse pages need fallback patterns; dense pages need progressive disclosure to avoid overwhelming.

Heading hierarchy that reflects data. H2s and H3s map to data sections (overview, details, comparisons, related). The heading structure is not decorative; it tells crawlers and AI engines what the page covers.

Schema markup as part of template. Structured data (JSON-LD, microdata, or RDFa depending on the stack) embedded in the template renders machine-readable signals at scale across the entire set. Without schema, the set is invisible to the structured-data extraction layer search and answer engines run.

Internal linking placeholders. The template includes link slots for related-record cross-references, parent-category links, and sibling-record links. 5 to 15 internal links per page is typical for a well-linked set.

Distinctive value per page. Each generated page must offer something the user could not get by going up to the parent or sideways to a sibling. If the page is just a re-pivot of the parent's data, the page is filler.

The template's quality bar. A randomly sampled page from the set, viewed in isolation, should answer the user's likely query competently. If a randomly sampled page is thin, the entire set is thin.

Detail in references/template-design-patterns.md.


Schema design

The data shape that drives the template. Design principles.

Field count signals depth. 5 fields per record is thin. 15 to 20 is competent. 30+ is deep. The field count is not a vanity metric; it determines what the template can render at depth versus what falls back to filler.

Required vs optional fields. Which fields must populate for a page to ship; templates need graceful degradation for optional gaps. A page with 3 of 30 optional fields populated should not ship; a page with 25 of 30 should.

Computed fields. Derived from source data. Average price per square foot computed from listings adds depth without requiring more source data. Computed fields are the pSEO equivalent of the writer adding analysis: the data does the analysis once, every page benefits.

Cross-record fields. Fields that reference other records in the set enable internal linking and comparison. The "5 most similar neighborhoods" field on every neighborhood page powers the sibling-link section without manual curation.

Update frequency tags. Which fields are static (geographic features, founding date) versus which need refresh tracking (current pricing, availability, current statistics). Without tagging, refresh becomes "audit everything" instead of "refresh the volatile fields."

The "schema-as-product" principle. The schema that drives pSEO IS the product surface. Treat it with the same rigor as a database schema for a customer-facing application. Versioned, reviewed, documented, breaking-change-aware.

Detail in references/schema-design-patterns.md.


Quality control at scale

Auditing all 100,000 pages individually is infeasible. The discipline.

Sampling strategy. Random sample 50 to 200 pages per audit cycle, balanced across data shape (sparse, dense, recent, old, popular categories, niche categories). The sample is not "the latest 50 pages"; it is a stratified sample that exposes problems specific to particular data shapes.

Automated checks. Heading structure, schema validity, internal link count, word count thresholds, duplicate-content checks, broken-link checks. Run on every page on a cadence; surface fail

Truncated for display — read the full file on GitHub.

Related Skills

View on GitHub
GitHub Stars902
CategoryMarketing
Updated2mo ago
Forks124

Languages

JavaScript

Trust signals

100/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

No cautions