Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Set up Talivia revenue-first website analytics with MCP, then verify live traffic and revenue attribution.
.claude/skills/hashgraph-online-talivia-agent-kit/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | -14% | 0% |
| case-23 | ✓→✓ | ＝ Same ✓ | 1% | 0% |
Talivia connects website traffic and visitor journeys to payment revenue. Use it when the user wants to install revenue-first analytics, discover which traffic becomes revenue, connect checkout attribution, or verify a Talivia setup.
talivia_account_status and talivia_websites_list.talivia_websites_create if creation was not explicit.talivia_tracking_snippet_get and talivia_framework_install_plan_get.talivia_tracker_verify and talivia_setup_status_get.talivia_payment_connect_start and send the user to the returned secure URL.talivia_payment_status_get and talivia_checkout_attribution_guide_get to finish verification.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 15,546 | 9,004 | -42% | 1 | 1 | 0% | 2,370 | 849 | -64% | 0 | 0 | — |
case-07 | fail→fail | 8,645 | 17,630 | +104% | 1 | 1 | 0% | 1,139 | 778 | -32% | 0 | 0 | — |
case-08 | fail→fail | 9,208 | 5,989 | -35% | 1 | 1 | 0% | 1,416 | 1,042 | -26% | 0 | 0 | — |
case-22 | pass→pass | 18,716 | 14,901 | -20% | 1 | 1 | 0% | 3,006 | 2,596 | -14% | 0 | 0 | — |
case-02 | fail→fail | 21,190 | 14,803 | -30% | 1 | 1 | 0% | 2,509 | 565 | -77% | 0 | 0 | — |
case-03 | fail→fail | 17,475 | 14,685 | -16% | 1 | 1 | 0% | 1,854 | 535 | -71% | 0 | 0 | — |
case-04 | fail→fail | 7,994 | 8,799 | +10% | 1 | 1 | 0% | 576 | 937 | +63% | 0 | 0 | — |
case-05 | fail→fail | 10,917 | 9,520 | -13% | 1 | 1 | 0% | 1,748 | 563 | -68% | 0 | 0 | — |
case-06 | fail→pass | 8,662 | 4,058 | -53% | 1 | 1 | 0% | 1,195 | 872 | -27% | 0 | 0 | — |
case-23 | pass→pass | 16,367 | 11,335 | -31% | 1 | 1 | 0% | 2,469 | 2,487 | +1% | 0 | 0 | — |
case-09 | fail→fail | 20,718 | 11,628 | -44% | 1 | 1 | 0% | 2,752 | 826 | -70% | 0 | 0 | — |
case-10 | fail→pass | 13,482 | 7,685 | -43% | 1 | 1 | 0% | 1,258 | 614 | -51% | 0 | 0 | — |
case-11 | fail→fail | 11,401 | 4,234 | -63% | 1 | 1 | 0% | 1,823 | 694 | -62% | 0 | 0 | — |
case-12 | fail→fail | 18,386 | 12,660 | -31% | 1 | 1 | 0% | 2,399 | 784 | -67% | 0 | 0 | — |
case-13 | fail→fail | 10,204 | 4,529 | -56% | 1 | 1 | 0% | 1,681 | 555 | -67% | 0 | 0 | — |
case-14 | fail→fail | 9,785 | 3,416 | -65% | 1 | 1 | 0% | 1,570 | 775 | -51% | 0 | 0 | — |
case-15 | pass→pass | 8,574 | 6,008 | -30% | 1 | 1 | 0% | 1,131 | 923 | -18% | 0 | 0 | — |
case-16 | fail→fail | 11,192 | 4,135 | -63% | 1 | 1 | 0% | 1,949 | 580 | -70% | 0 | 0 | — |
case-17 | fail→pass | 6,425 | 3,725 | -42% | 1 | 1 | 0% | 1,032 | 771 | -25% | 0 | 0 | — |
case-18 | fail→fail | 20,708 | 4,919 | -76% | 1 | 1 | 0% | 3,960 | 280 | -93% | 0 | 0 | — |
case-19 | fail→fail | 8,548 | 8,340 | -2% | 1 | 1 | 0% | 1,270 | 835 | -34% | 0 | 0 | — |
case-20 | pass→pass | 14,344 | 11,079 | -23% | 1 | 1 | 0% | 2,140 | 1,879 | -12% | 0 | 0 | — |
case-21 | pass→pass | 15,179 | 15,009 | -1% | 1 | 1 | 0% | 2,313 | 2,538 | +10% | 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, and 14 counted toward the lift figure. The other 9 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +13 percentage points is the difference between those two pass rates over the 14 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.