Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Navigate Taiwan fintech regulations including FSC oversight, electronic payment laws, VASP rules, AML/KYC requirements, and the regulatory sandbox. Use this skill when the user is building a fintech product in Taiwan, needs to understand licensing requirements, assess crypto/VASP compliance, or apply for the regulatory sandbox — even if they say 'do we need a license', 'crypto regulation in Taiwan', 'KYC requirements', or 'fintech sandbox application'.
.claude/skills/asgard-ai-platform-tw-fintech-compliance/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 9% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -8% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 22% | 0% |
IRON LAW: In Fintech, Regulation Comes BEFORE Product
Unlike most tech products where you can launch and iterate, financial
services in Taiwan require LICENSING FIRST. Operating without the proper
license is a criminal offense (not just a fine). Determine your regulatory
category and licensing requirements BEFORE writing code.金融監督管理委員會 (FSC) oversees all financial services in Taiwan:
| Category | Governing Law | License Required | FSC Division | |----------|-------------|-----------------|-------------| | Electronic payment | 電子支付機構管理條例 | 電子支付機構許可 | Banking Bureau | | Third-party payment | 第三方支付服務業管理辦法 | Registration (lighter) | Banking Bureau | | P2P lending | 金融科技發展與創新實驗條例 (sandbox) or crowd-funding rules | Case-by-case | Securities Bureau | | Robo-advisory | 證券投資信託及顧問法 | Investment advisory license | Securities Bureau | | Insurance tech | 保險法 | Insurance broker/agent license | Insurance Bureau | | VASP (crypto) | 虛擬資產服務提供者管理辦法 (2024) | VASP registration + AML compliance | Securities Bureau | | Open banking | 金融機構間資料共享指引 | API partnership with banks | Banking Bureau |
| Feature | Electronic Payment (電子支付) | Third-Party Payment (第三方支付) | |---------|---------------------------|-------------------------------| | Stored value | Yes (e-wallet) | No (pass-through only) | | P2P transfer | Yes | No | | Capital requirement | NT$500M+ | NT$5M | | License | FSC approval (12-18 months) | Registration (1-3 months) | | Examples | LINE Pay Money, JKoPay, icash Pay | 綠界 ECPay, 藍新 NewebPay |
| Requirement | Detail | |------------|--------| | Registration | Must register with FSC as VASP | | AML/CFT | Full compliance with 洗錢防制法 | | KYC | Identity verification for all users | | Custody | Customer asset segregation required | | Marketing | Restrictions on advertising (no promises of returns) | | Reporting | Suspicious transaction reports (STR) to 調查局 |
| KYC Level | When Required | Data Collected | |-----------|-------------|---------------| | Simplified | Low-risk, small transactions | Name, ID number, DOB | | Standard | Account opening, most transactions | + Address, occupation, source of funds | | Enhanced | High-risk customers, PEPs, large transactions | + Detailed source of wealth, ongoing monitoring |
Ongoing obligations:
| Aspect | Detail | |--------|--------| | Purpose | Test innovative financial services without full licensing | | Duration | Up to 18 months (extendable to 36) | | Application | Submit to FSC with: innovation description, consumer protection plan, risk assessment | | Approval time | 60 business days | | Scope limitation | Must define user count, transaction limits, geographic scope | | Exit plan | Path to full license or orderly shutdown |
| Phase | Data Shared | Status | |-------|-----------|--------| | Phase 1 | Product information (rates, fees) | Launched 2019 | | Phase 2 | Customer information (with consent) | Launched 2021 | | Phase 3 | Transaction initiation (payment, transfer) | In development |
markdown# Fintech Compliance Assessment: {Product} ## Product Classification - Service type: {payment / lending / advisory / crypto / insurance} - Regulatory category: {electronic payment / third-party / VASP / etc.} - License required: {specific license name} ## Licensing Pathway | Step | Action | Timeline | Cost | |------|--------|----------|------| | 1 | {step} | {months} | NT${X} | ## AML/KYC Requirements - KYC level: {simplified / standard / enhanced} - Transaction monitoring: {required / not required} - STR reporting: {required / not required} ## Key Risks | Risk | Impact | Mitigation | |------|--------|-----------| | {regulatory risk} | H/M/L | {action} | ## Recommendation - Proceed with licensing / Use sandbox / Restructure product to avoid licensing
references/fsc-licensing.mdreferences/aml-kyc-checklist.md| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 30,882 | 32,474 | +5% | 1 | 1 | 0% | 5,335 | 4,922 | -8% | 0 | 0 | — |
case-02 | fail→pass | 23,515 | 19,353 | -18% | 1 | 1 | 0% | 3,849 | 4,062 | +6% | 0 | 0 | — |
case-03 | fail→pass | 28,697 | 18,246 | -36% | 1 | 1 | 0% | 4,107 | 4,484 | +9% | 0 | 0 | — |
case-04 | fail→pass | 24,251 | 18,577 | -23% | 1 | 1 | 0% | 3,744 | 4,198 | +12% | 0 | 0 | — |
case-05 | pass→pass | 29,066 | 18,872 | -35% | 1 | 1 | 0% | 4,630 | 4,596 | -1% | 0 | 0 | — |
case-06 | fail→pass | 29,093 | 21,538 | -26% | 1 | 1 | 0% | 4,705 | 4,345 | -8% | 0 | 0 | — |
case-07 | pass→pass | 22,469 | 17,482 | -22% | 1 | 1 | 0% | 3,429 | 4,149 | +21% | 0 | 0 | — |
case-08 | fail→fail | 17,969 | 12,470 | -31% | 1 | 1 | 0% | 2,780 | 3,401 | +22% | 0 | 0 | — |
case-09 | pass→pass | 21,197 | 15,867 | -25% | 1 | 1 | 0% | 3,125 | 3,996 | +28% | 0 | 0 | — |
case-10 | pass→pass | 19,518 | 16,014 | -18% | 1 | 1 | 0% | 3,077 | 3,907 | +27% | 0 | 0 | — |
case-11 | pass→pass | 27,024 | 19,716 | -27% | 1 | 1 | 0% | 4,168 | 4,460 | +7% | 0 | 0 | — |
case-12 | pass→pass | 25,368 | 22,583 | -11% | 1 | 1 | 0% | 3,427 | 4,436 | +29% | 0 | 0 | — |
case-13 | fail→pass | 24,273 | 15,844 | -35% | 1 | 1 | 0% | 3,248 | 3,966 | +22% | 0 | 0 | — |
case-14 | pass→pass | 22,320 | 21,788 | -2% | 1 | 1 | 0% | 3,397 | 4,884 | +44% | 0 | 0 | — |
case-15 | pass→pass | 23,840 | 21,434 | -10% | 1 | 1 | 0% | 3,728 | 4,296 | +15% | 0 | 0 | — |
case-16 | fail→pass | 28,132 | 21,867 | -22% | 1 | 1 | 0% | 4,136 | 4,140 | +0% | 0 | 0 | — |
case-17 | fail→pass | 22,656 | 14,895 | -34% | 1 | 1 | 0% | 3,539 | 3,850 | +9% | 0 | 0 | — |
case-18 | fail→pass | 19,462 | 22,020 | +13% | 1 | 1 | 0% | 3,253 | 4,782 | +47% | 0 | 0 | — |
case-19 | pass→pass | 25,669 | 15,866 | -38% | 1 | 1 | 0% | 3,471 | 4,119 | +19% | 0 | 0 | — |
case-20 | fail→pass | 24,808 | 21,759 | -12% | 1 | 1 | 0% | 3,375 | 4,407 | +31% | 0 | 0 | — |
case-21 | pass→pass | 22,856 | 12,688 | -44% | 1 | 1 | 0% | 3,155 | 3,409 | +8% | 0 | 0 | — |
case-22 | pass→pass | 25,174 | 22,831 | -9% | 1 | 1 | 0% | 3,460 | 5,004 | +45% | 0 | 0 | — |
case-23 | pass→pass | 29,271 | 21,002 | -28% | 1 | 1 | 0% | 3,856 | 4,688 | +22% | 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 +39 percentage points is the difference between those two pass rates over the 23 comparable cases.
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.