Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Integrate TapPay for Taiwan e-commerce — TapPay Web/iOS/Android SDK, 3DS 2.0 flow, Apple Pay / Google Pay / LINE Pay integration via TapPay, TCS (TapPay Card Service) for tokenized recurring payments. Use when TapPay is chosen as PSP, comparing TapPay vs NewebPay, or debugging 3DS flows. Do NOT use for other gateways. STATUS: SKELETON — body pending.
.claude/skills/asgard-ai-platform-tw-ecom-payment-tappay/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-01 | ✓→✓ | ＝ Same ✓ | 1% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 6% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 30% | 0% |
> STATUS: SKELETON — body pending.
tw-payment-integrationTODO: TapPay positioning (mobile-first), TapPay Direct vs TapPay Pay by Prime, TCS flow.
TODO: when TapPay outperforms NewebPay/ECPay.
TODO: Prime token acquisition, 3DS handling, TCS card binding + charge.
TODO: 5-6 pitfalls (Prime token TTL, 3DS interstitial UX, TCS unbind on card expiry, Apple Pay domain verification, sandbox-to-prod merchant switch).
TODO.
TODO.
tw-payment-integration for gateway comparisontw-ecom-payment-dispute for refund / chargeback handling_Last verified: 2026-04_
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 45,596 | 45,394 | -0% | 1 | 1 | 0% | 2,483 | 2,516 | +1% | 0 | 0 | — |
case-02 | fail→pass | 48,756 | 12,954 | -73% | 1 | 1 | 0% | 1,701 | 2,605 | +53% | 0 | 0 | — |
case-03 | pass→pass | 11,963 | 23,626 | +97% | 1 | 1 | 0% | 2,406 | 2,546 | +6% | 0 | 0 | — |
case-04 | fail→pass | 8,360 | 38,778 | +364% | 1 | 1 | 0% | 1,215 | 1,543 | +27% | 0 | 0 | — |
case-05 | pass→pass | 11,454 | 44,186 | +286% | 1 | 1 | 0% | 2,186 | 2,838 | +30% | 0 | 0 | — |
case-06 | pass→pass | 44,904 | 12,592 | -72% | 1 | 1 | 0% | 2,117 | 2,662 | +26% | 0 | 0 | — |
case-07 | pass→pass | 10,572 | 10,988 | +4% | 1 | 1 | 0% | 1,809 | 2,403 | +33% | 0 | 0 | — |
case-08 | pass→pass | 13,561 | 27,837 | +105% | 1 | 1 | 0% | 2,549 | 2,651 | +4% | 0 | 0 | — |
case-09 | pass→pass | 16,197 | 22,168 | +37% | 1 | 1 | 0% | 2,780 | 3,653 | +31% | 0 | 0 | — |
case-10 | pass→pass | 20,189 | 16,724 | -17% | 1 | 1 | 0% | 2,696 | 3,213 | +19% | 0 | 0 | — |
case-11 | pass→pass | 14,270 | 22,325 | +56% | 1 | 1 | 0% | 2,463 | 3,469 | +41% | 0 | 0 | — |
case-12 | pass→pass | 16,816 | 16,144 | -4% | 1 | 1 | 0% | 3,028 | 3,317 | +10% | 0 | 0 | — |
case-13 | pass→pass | 18,295 | 22,651 | +24% | 1 | 1 | 0% | 2,900 | 3,446 | +19% | 0 | 0 | — |
case-14 | fail→fail | 15,001 | 17,884 | +19% | 1 | 1 | 0% | 2,684 | 2,918 | +9% | 0 | 0 | — |
case-15 | pass→pass | 9,305 | 6,438 | -31% | 1 | 1 | 0% | 1,335 | 1,564 | +17% | 0 | 0 | — |
case-16 | pass→pass | 10,576 | 13,714 | +30% | 1 | 1 | 0% | 2,027 | 2,872 | +42% | 0 | 0 | — |
case-17 | fail→fail | 16,150 | 14,000 | -13% | 1 | 1 | 0% | 2,411 | 2,698 | +12% | 0 | 0 | — |
case-18 | pass→pass | 13,286 | 14,593 | +10% | 1 | 1 | 0% | 2,333 | 2,506 | +7% | 0 | 0 | — |
case-19 | pass→pass | 17,976 | 32,313 | +80% | 1 | 1 | 0% | 3,353 | 6,397 | +91% | 0 | 0 | — |
case-20 | pass→pass | 7,606 | 7,684 | +1% | 1 | 1 | 0% | 1,100 | 1,786 | +62% | 0 | 0 | — |
case-21 | pass→pass | 16,314 | 14,320 | -12% | 1 | 1 | 0% | 2,758 | 2,697 | -2% | 0 | 0 | — |
case-22 | pass→pass | 13,498 | 13,426 | -1% | 1 | 1 | 0% | 2,411 | 2,731 | +13% | 0 | 0 | — |
case-23 | pass→pass | 18,509 | 18,679 | +1% | 1 | 1 | 0% | 2,999 | 3,301 | +10% | 0 | 0 | — |
case-24 | pass→pass | 20,660 | 25,238 | +22% | 1 | 1 | 0% | 3,883 | 5,381 | +39% | 0 | 0 | — |
case-25 | pass→pass | 19,877 | 18,082 | -9% | 1 | 1 | 0% | 2,901 | 3,061 | +6% | 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. 25 cases were attempted. The headline lift of +8 percentage points is the difference between those two pass rates over the 25 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.