Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Structure a dev-tool homepage that converts developers into champions. Use when the landing page is written for the buyer instead of the developer, reads as salesy, buries what the product does, or makes it hard to start. Pairs with the ShipReady homepage audit.
.claude/skills/aidevgtm-the-homepage/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -35% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 19% | 0% |
> Write your homepage for the person who actually reads it: the end-user developer, not the CTO who signs the check. The developer becomes your champion and sells it internally for you.
Use this when: your homepage speaks to buyers/ROI, it's clever but vague about what the thing does, or a developer can't tell in 5 seconds whether to care, and can't start in 5 minutes if they do.
Address end-user developers on the homepage; put enterprise/ROI/buyer messaging on separate pages (enterprise, pricing, case studies). If you write the homepage for the buyer, you've written it for someone who isn't there.
value-prop-that-converts).Design: clean typography, breathing room, no "salesy BS," minimal heavy interactions. An embedded playground / copy-pasteable snippet creates an instant aha.
Cross-check the copy has all six:
Don't build everything. Start with: homepage · docs · pricing · navbar (login, signup, contact, GitHub) · About (an About page builds trust for an unknown brand). Add later, deliberately: case studies → comparison pages → blog → templates → sandbox → enterprise page.
> Transparent pricing. "Contact sales for pricing" reads as "expensive and slow." Show a number wherever you can.
Can a developer get value self-serve?
├─ YES → primary CTA "Get started" (signup/download) + secondary "Docs".
└─ NO → primary CTA "Read the docs" / "See how it works"; don't fake a signup that dead-ends.value-prop-that-converts).Built from real dev-tool GTM experience, with frameworks from Adam Frankl (The Developer-Facing Startup) and Jakub Czakon (markepear.dev). When a framework can't make the call, that's what a human is for: The DevTool GTM Company.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 21,345 | 20,828 | -2% | 1 | 1 | 0% | 3,267 | 4,287 | +31% | 0 | 0 | — |
case-02 | fail→fail | 20,865 | 17,761 | -15% | 1 | 1 | 0% | 3,122 | 3,454 | +11% | 0 | 0 | — |
case-03 | fail→pass | 40,838 | 19,107 | -53% | 1 | 1 | 0% | 6,114 | 3,979 | -35% | 0 | 0 | — |
case-04 | pass→pass | 22,692 | 18,919 | -17% | 1 | 1 | 0% | 3,254 | 3,778 | +16% | 0 | 0 | — |
case-05 | pass→pass | 14,328 | 16,310 | +14% | 1 | 1 | 0% | 2,408 | 3,636 | +51% | 0 | 0 | — |
case-06 | pass→pass | 29,510 | 36,072 | +22% | 1 | 1 | 0% | 4,898 | 7,348 | +50% | 0 | 0 | — |
case-07 | fail→pass | 14,801 | 12,582 | -15% | 1 | 1 | 0% | 2,249 | 2,714 | +21% | 0 | 0 | — |
case-08 | fail→fail | 14,170 | 12,835 | -9% | 1 | 1 | 0% | 2,013 | 2,795 | +39% | 0 | 0 | — |
case-09 | fail→pass | 17,230 | 14,229 | -17% | 1 | 1 | 0% | 2,447 | 2,956 | +21% | 0 | 0 | — |
case-10 | fail→pass | 16,057 | 13,208 | -18% | 1 | 1 | 0% | 2,502 | 2,886 | +15% | 0 | 0 | — |
case-11 | pass→pass | 14,759 | 12,194 | -17% | 1 | 1 | 0% | 2,151 | 2,780 | +29% | 0 | 0 | — |
case-12 | pass→pass | 17,665 | 14,611 | -17% | 1 | 1 | 0% | 2,340 | 3,068 | +31% | 0 | 0 | — |
case-13 | pass→pass | 17,840 | 15,227 | -15% | 1 | 1 | 0% | 2,646 | 3,116 | +18% | 0 | 0 | — |
case-14 | pass→pass | 15,659 | 13,520 | -14% | 1 | 1 | 0% | 2,175 | 2,919 | +34% | 0 | 0 | — |
case-15 | pass→fail | 13,753 | 11,332 | -18% | 1 | 1 | 0% | 1,960 | 2,580 | +32% | 0 | 0 | — |
case-16 | pass→pass | 17,341 | 16,149 | -7% | 1 | 1 | 0% | 2,444 | 3,297 | +35% | 0 | 0 | — |
case-17 | fail→fail | 14,009 | 11,862 | -15% | 1 | 1 | 0% | 2,195 | 2,379 | +8% | 0 | 0 | — |
case-18 | fail→fail | 15,327 | 10,109 | -34% | 1 | 1 | 0% | 2,222 | 2,261 | +2% | 0 | 0 | — |
case-19 | pass→pass | 14,582 | 14,020 | -4% | 1 | 1 | 0% | 2,062 | 2,993 | +45% | 0 | 0 | — |
case-20 | fail→pass | 16,776 | 12,153 | -28% | 1 | 1 | 0% | 2,275 | 2,708 | +19% | 0 | 0 | — |
case-21 | fail→pass | 11,614 | 9,763 | -16% | 1 | 1 | 0% | 1,763 | 2,623 | +49% | 0 | 0 | — |
case-22 | fail→pass | 15,969 | 13,242 | -17% | 1 | 1 | 0% | 2,290 | 2,704 | +18% | 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 +27 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is 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.