Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Design cross-border logistics strategies including direct mail, overseas warehousing, and bonded warehouse models for international e-commerce. Use this skill when the user needs to ship products internationally, choose a logistics model for cross-border sales, optimize shipping costs, or set up fulfillment in a foreign market — even if they say 'ship to Southeast Asia', 'overseas warehouse vs direct shipping', 'customs clearance', or 'reduce international shipping time'.
.claude/skills/asgard-ai-platform-xborder-logistics/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 41% | 0% |
IRON LAW: Return Logistics Costs 3-5× More Than Outbound — Budget for It or Die
Agents plan outbound shipping costs but systematically ignore returns.
Cross-border return rates run 15-30% for apparel, and each return costs
3-5× the outbound shipment (reverse customs, restocking, re-export duties,
currency loss). If your margin can't absorb a 20% return rate at 4× cost,
the overseas warehouse model is a cash trap regardless of forward volume.
Calculate the break-even return rate BEFORE committing to a fulfillment model.| Model | How It Works | Delivery Time | Cost per Order | Min. Volume | Best For | |-------|-------------|-------------|---------------|------------|----------| | Direct Mail (直郵) | Ship each order from Taiwan to customer | 7-21 days | High ($8-25) | 1 order | Testing market, low volume, high-value items | | Overseas Warehouse (海外倉) | Pre-stock inventory in destination country, ship locally | 1-5 days | Low ($2-5 local) + warehousing | 100+ units/month | Proven demand, competitive delivery needed | | Bonded Warehouse (保稅倉) | Store in bonded zone, clear customs per-order | 3-7 days | Medium ($5-10) | 50+ units/month | Duty deferral, uncertain demand |
Monthly orders to one country < 50 → Direct Mail
Monthly orders 50-500 → Consider Bonded Warehouse
Monthly orders > 500 → Overseas Warehouse justified| Factor | What to Know | |--------|-------------| | De minimis threshold | Below this value, no import duty (varies: $75 in most SEA, $400 in US) | | HS Code | Product classification code determines duty rate. Get this RIGHT — wrong HS code = penalties | | Landed cost | Product cost + shipping + insurance + duties + taxes = what customer actually pays | | Documentation | Commercial invoice, packing list, certificate of origin, product-specific certificates | | Restricted items | Food, cosmetics, electronics, medical devices often need import permits |
| Type | Service | Cost | Control | |------|---------|------|---------| | 3PL (e.g., ShipBob, Boxme) | Full service: storage, pick-pack, ship | $$-$$$ | Low (outsourced) | | Marketplace fulfillment (Shopee/Lazada warehouse) | Platform handles logistics | Commission-included | Very low | | Self-operated warehouse | Your own warehouse + staff | $$$$ upfront | Full | | Drop-shipping | Supplier ships directly | Lowest | None |
Phase 1: Direct Mail (Market Testing)
Phase 2: Transition to Overseas Warehouse (Scaling)
Phase 3: Optimize
markdown# Cross-Border Logistics Plan: {Product} → {Destination} ## Current State - Monthly orders: {N} - Current model: {direct mail / none} - Avg delivery time: {days} ## Recommended Model - Model: {Direct Mail / Bonded / Overseas Warehouse} - Rationale: {volume, speed requirement, cost} ## Cost Comparison | Model | Per-Order Cost | Monthly Fixed | Total (at {N} orders) | |-------|---------------|-------------|---------------------| | Direct Mail | ${X} | $0 | ${X} | | Overseas Warehouse | ${X} | ${X} | ${X} | ## Implementation Plan | Phase | Action | Timeline | |-------|--------|----------| | 1 | {step} | {weeks} |
references/customs-by-country.mdreferences/3pl-comparison.md| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 26,615 | 70,111 | +163% | 1 | 1 | 0% | 3,790 | 5,181 | +37% | 0 | 0 | — |
case-02 | fail→pass | 33,436 | 51,601 | +54% | 1 | 1 | 0% | 4,982 | 4,397 | -12% | 0 | 0 | — |
case-03 | fail→pass | 33,478 | 23,272 | -30% | 1 | 1 | 0% | 4,563 | 5,004 | +10% | 0 | 0 | — |
case-04 | pass→pass | 28,402 | 22,549 | -21% | 1 | 1 | 0% | 4,954 | 5,010 | +1% | 0 | 0 | — |
case-05 | pass→pass | 52,413 | 35,859 | -32% | 1 | 1 | 0% | 8,273 | 6,065 | -27% | 0 | 0 | — |
case-06 | pass→pass | 15,280 | 17,905 | +17% | 1 | 1 | 0% | 2,484 | 3,566 | +44% | 0 | 0 | — |
case-07 | pass→pass | 33,182 | 24,325 | -27% | 1 | 1 | 0% | 4,592 | 4,952 | +8% | 0 | 0 | — |
case-08 | fail→fail | 37,390 | 29,954 | -20% | 1 | 1 | 0% | 4,943 | 5,199 | +5% | 0 | 0 | — |
case-09 | fail→pass | 26,277 | 24,462 | -7% | 1 | 1 | 0% | 4,317 | 5,239 | +21% | 0 | 0 | — |
case-10 | pass→pass | 26,333 | 23,061 | -12% | 1 | 1 | 0% | 4,113 | 5,121 | +25% | 0 | 0 | — |
case-11 | fail→fail | 21,950 | 20,998 | -4% | 1 | 1 | 0% | 2,648 | 4,624 | +75% | 0 | 0 | — |
case-12 | fail→fail | 30,553 | 27,276 | -11% | 1 | 1 | 0% | 4,074 | 5,741 | +41% | 0 | 0 | — |
case-13 | pass→pass | 22,440 | 22,357 | -0% | 1 | 1 | 0% | 3,129 | 4,823 | +54% | 0 | 0 | — |
case-14 | pass→pass | 15,147 | 23,545 | +55% | 1 | 1 | 0% | 2,477 | 4,507 | +82% | 0 | 0 | — |
case-15 | pass→fail | 33,334 | 26,264 | -21% | 1 | 1 | 0% | 4,479 | 5,483 | +22% | 0 | 0 | — |
case-16 | pass→pass | 21,668 | 30,782 | +42% | 1 | 1 | 0% | 3,367 | 5,454 | +62% | 0 | 0 | — |
case-17 | fail→pass | 21,451 | 20,505 | -4% | 1 | 1 | 0% | 3,325 | 4,449 | +34% | 0 | 0 | — |
case-18 | pass→pass | 23,467 | 23,425 | -0% | 1 | 1 | 0% | 3,801 | 5,114 | +35% | 0 | 0 | — |
case-19 | pass→pass | 15,733 | 22,434 | +43% | 1 | 1 | 0% | 2,432 | 4,304 | +77% | 0 | 0 | — |
case-20 | fail→pass | 19,245 | 25,071 | +30% | 1 | 1 | 0% | 3,122 | 4,391 | +41% | 0 | 0 | — |
case-21 | fail→fail | 13,711 | 12,735 | -7% | 1 | 1 | 0% | 2,218 | 3,117 | +41% | 0 | 0 | — |
case-22 | fail→pass | 19,861 | 37,011 | +86% | 1 | 1 | 0% | 3,359 | 4,829 | +44% | 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 +23 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.