Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Plan a programmatic SEO strategy — generate many ranking pages from a data set and a template. Use when asked about pSEO, scaling content with templates/data, building [X] for [Y] pages, or capturing long-tail search at scale. Produces the head-term + modifier model, the page template and data schema, a quality/thin-content guardrail, and an indexation plan — pages worth ranking, not doorway spam.
.claude/skills/mohitagw15856-programmatic-seo/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 36% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -11% | 0% |
Programmatic SEO turns a data set + a template into hundreds or thousands of pages that each target a specific long-tail query ("best tool] for use case]", "city] service]"). Done well it captures huge long-tail demand; done badly it's thin doorway spam that gets deindexed. This skill plans the good version — real data, real value per page, and the guardrails to stay on the right side of that line.
Ask for these only if they aren't already provided:
[integration] + alternatives, [role] + templates).1. The page model — head term × modifier(s) → URL pattern, and the realistic page count. Prioritise the modifier sets with real search volume and commercial/informational intent.
2. Page template — the sections every page has, and what makes each page genuinely useful (unique data, comparisons, specifics) — not just swapped keywords. Show the template with data placeholders.
3. Data schema — the fields each page needs, the source, and how it stays fresh. (No data = thin page.)
4. Quality guardrail — the bar a page must clear to be published (enough unique value, real data, intent match). Pages that can't clear it shouldn't exist. How to avoid near-duplicate/thin pages.
5. Internal linking & indexation — hub/spoke linking, sitemaps, and a phased rollout (publish a quality batch, confirm it indexes and ranks, then scale) rather than dumping 5,000 pages day one.
6. Measurement — what to watch (indexed %, rankings, traffic, conversion) and the kill criterion for pages that never rank.
Programmatic SEO practice (templated data-driven pages, intent + unique value, Google's thin-content/helpful-content guidance).
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-15 | fail→fail | 30,645 | 41,888 | +37% | 1 | 1 | 0% | 3,967 | 5,562 | +40% | 0 | 0 | — |
case-01 | fail→fail | 60,199 | 63,284 | +5% | 1 | 1 | 0% | 8,243 | 6,376 | -23% | 0 | 0 | — |
case-02 | fail→pass | 52,170 | 40,255 | -23% | 1 | 1 | 0% | 8,288 | 5,671 | -32% | 0 | 0 | — |
case-03 | fail→fail | 48,561 | 49,440 | +2% | 1 | 1 | 0% | 8,290 | 6,893 | -17% | 0 | 0 | — |
case-04 | fail→pass | 26,953 | 26,831 | -0% | 1 | 1 | 0% | 3,511 | 4,064 | +16% | 0 | 0 | — |
case-05 | fail→pass | 31,780 | 29,422 | -7% | 1 | 1 | 0% | 3,695 | 5,013 | +36% | 0 | 0 | — |
case-06 | fail→fail | 27,770 | 30,936 | +11% | 1 | 1 | 0% | 3,641 | 4,424 | +22% | 0 | 0 | — |
case-07 | fail→fail | 30,417 | 29,157 | -4% | 1 | 1 | 0% | 3,278 | 4,531 | +38% | 0 | 0 | — |
case-08 | fail→fail | 42,576 | 30,843 | -28% | 1 | 1 | 0% | 3,831 | 4,782 | +25% | 0 | 0 | — |
case-09 | fail→fail | 31,778 | 25,817 | -19% | 1 | 1 | 0% | 3,572 | 4,722 | +32% | 0 | 0 | — |
case-10 | pass→pass | 26,676 | 39,942 | +50% | 1 | 1 | 0% | 4,160 | 4,912 | +18% | 0 | 0 | — |
case-11 | fail→fail | 25,671 | 32,712 | +27% | 1 | 1 | 0% | 3,789 | 5,083 | +34% | 0 | 0 | — |
case-12 | pass→pass | 23,702 | 28,911 | +22% | 1 | 1 | 0% | 3,149 | 4,358 | +38% | 0 | 0 | — |
case-13 | pass→fail | 28,440 | 32,861 | +16% | 1 | 1 | 0% | 4,075 | 5,231 | +28% | 0 | 0 | — |
case-14 | pass→pass | 31,138 | 31,624 | +2% | 1 | 1 | 0% | 4,279 | 6,127 | +43% | 0 | 0 | — |
case-16 | fail→pass | 25,466 | 28,655 | +13% | 1 | 1 | 0% | 4,176 | 4,647 | +11% | 0 | 0 | — |
case-17 | pass→pass | 25,747 | 22,869 | -11% | 1 | 1 | 0% | 3,948 | 4,486 | +14% | 0 | 0 | — |
case-18 | fail→pass | 25,867 | 19,092 | -26% | 1 | 1 | 0% | 3,943 | 3,518 | -11% | 0 | 0 | — |
case-19 | fail→pass | 32,150 | 47,688 | +48% | 1 | 1 | 0% | 4,705 | 4,646 | -1% | 0 | 0 | — |
case-20 | pass→pass | 23,557 | 21,061 | -11% | 1 | 1 | 0% | 3,551 | 4,024 | +13% | 0 | 0 | — |
case-21 | pass→pass | 17,635 | 17,731 | +1% | 1 | 1 | 0% | 2,316 | 2,927 | +26% | 0 | 0 | — |
case-22 | pass→pass | 23,817 | 30,878 | +30% | 1 | 1 | 0% | 3,270 | 4,969 | +52% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +23 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.