Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create and optimize popups, modals, overlays, slide-ins, and banners to increase conversions without harming user experience or brand trust.
.claude/skills/dokhacgiakhoa-popup-cro/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -55% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 70% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 71% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 23% | 0% |
You are an expert in popup and modal optimization. Your goal is to design high-converting, respectful interruption patterns that capture value at the right moment—without annoying users, harming trust, or violating SEO or accessibility guidelines.
This skill focuses on strategy, copy, triggers, and rules. For optimizing the form inside the popup, see form-cro. For optimizing the page itself, see page-cro.
Before making recommendations, establish context:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 30,559 | 25,276 | -17% | 1 | 1 | 0% | 4,833 | 4,862 | +1% | 0 | 0 | — |
case-02 | fail→fail | 17,028 | 13,919 | -18% | 1 | 1 | 0% | 2,860 | 3,209 | +12% | 0 | 0 | — |
case-03 | fail→fail | 17,969 | 21,101 | +17% | 1 | 1 | 0% | 2,974 | 4,349 | +46% | 0 | 0 | — |
case-04 | fail→pass | 23,485 | 8,173 | -65% | 1 | 1 | 0% | 4,950 | 2,227 | -55% | 0 | 0 | — |
case-05 | fail→pass | 15,385 | 8,449 | -45% | 1 | 1 | 0% | 2,590 | 2,315 | -11% | 0 | 0 | — |
case-06 | fail→fail | 16,337 | 18,197 | +11% | 1 | 1 | 0% | 3,339 | 4,499 | +35% | 0 | 0 | — |
case-07 | pass→pass | 11,610 | 12,010 | +3% | 1 | 1 | 0% | 1,800 | 2,850 | +58% | 0 | 0 | — |
case-08 | fail→pass | 14,037 | 17,018 | +21% | 1 | 1 | 0% | 2,139 | 3,633 | +70% | 0 | 0 | — |
case-09 | pass→pass | 10,321 | 13,749 | +33% | 1 | 1 | 0% | 1,719 | 3,246 | +89% | 0 | 0 | — |
case-10 | pass→pass | 11,052 | 11,026 | -0% | 1 | 1 | 0% | 1,775 | 2,669 | +50% | 0 | 0 | — |
case-11 | fail→pass | 11,907 | 14,407 | +21% | 1 | 1 | 0% | 1,877 | 3,206 | +71% | 0 | 0 | — |
case-12 | fail→pass | 14,787 | 12,440 | -16% | 1 | 1 | 0% | 2,450 | 3,024 | +23% | 0 | 0 | — |
case-13 | pass→pass | 12,650 | 13,584 | +7% | 1 | 1 | 0% | 2,122 | 3,282 | +55% | 0 | 0 | — |
case-14 | pass→pass | 11,274 | 11,820 | +5% | 1 | 1 | 0% | 1,738 | 2,781 | +60% | 0 | 0 | — |
case-15 | pass→pass | 12,834 | 10,932 | -15% | 1 | 1 | 0% | 2,088 | 2,608 | +25% | 0 | 0 | — |
case-16 | pass→pass | 11,902 | 14,441 | +21% | 1 | 1 | 0% | 1,915 | 3,299 | +72% | 0 | 0 | — |
case-17 | pass→fail | 10,215 | 9,445 | -8% | 1 | 1 | 0% | 1,629 | 2,428 | +49% | 0 | 0 | — |
case-18 | pass→pass | 10,106 | 8,953 | -11% | 1 | 1 | 0% | 1,619 | 2,354 | +45% | 0 | 0 | — |
case-19 | pass→pass | 10,956 | 11,576 | +6% | 1 | 1 | 0% | 1,728 | 2,521 | +46% | 0 | 0 | — |
case-20 | pass→pass | 14,486 | 12,162 | -16% | 1 | 1 | 0% | 2,170 | 2,837 | +31% | 0 | 0 | — |
case-21 | pass→pass | 12,754 | 16,083 | +26% | 1 | 1 | 0% | 2,163 | 3,212 | +48% | 0 | 0 | — |
case-22 | pass→pass | 10,011 | 8,893 | -11% | 1 | 1 | 0% | 1,612 | 2,288 | +42% | 0 | 0 | — |
case-23 | pass→pass | 13,611 | 16,558 | +22% | 1 | 1 | 0% | 2,456 | 3,252 | +32% | 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. 23 cases were attempted. The headline lift of +17 percentage points is the difference between those two pass rates over the 23 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.