Menu
Multi-channel e-commerce operations — Shopify, Amazon, eBay order management, inventory sync, returns, fulfillment, and P&L reports. Use when the user mentions Shopify, Amazon, eBay, orders, inventory, returns, shipping, SKUs, listings, or sales reports.
Travel Agent — book AND cancel real hotels and stays end to end. Use when the user wants to BOOK or RESERVE a hotel/stay (not just look one up): "book me a hotel in Lisbon", "reserve that room"; or to CANCEL / look up a booking they made through you: "cancel my hotel", "what's my booking". Discovery is free; booking spends real USDC and always requires the user's explicit confirmation; cancelling requires a code Travala emails to the booking address. Hotels and stays only (no flights). Routes payment through the platform's proven x402 rail — never invent payment logic, never sign a transfer yourself.
Your economic toolbelt. Use when the user mentions earning, selling, charging, invoicing, paying, spending, x402, stripe, reputation, agent economy, offerings, marketplace, transactions, or asks "can you make money" / "how do I earn". Uses open standards (x402, Stripe MCP, AP2, ERC-8004) — never invent payment logic; always use the official tools listed below.
Extend or edit an EXISTING video the user already has (lengthen/continue a clip, multi-shot stitch of existing footage, audio for a video they already made). NOT for new videos or images — for any NEW video, clip, animation, or image use the higgsfield-cloud skill, and never route here as a fallback when higgsfield-cloud is blocked.
Swap tokens, provide liquidity, and stake on Aerodrome (Base mainnet) — Base's leading DEX, Velodrome v2 fork with stable + volatile pools.
Supply, borrow, and manage positions on Moonwell (Base mainnet) — a Compound v2-style money market with mUSDC, mWETH, mcbETH, mcbBTC.
Swap tokens and manage liquidity on Uniswap v3 (Base mainnet) — Universal Router for swaps, NonfungiblePositionManager for v3 positions.
| name | ecommerce-marketplace |
| description | Multi-channel e-commerce operations — Shopify, Amazon, eBay order management, inventory sync, returns, fulfillment, and P&L reports. Use when the user mentions Shopify, Amazon, eBay, orders, inventory, returns, shipping, SKUs, listings, or sales reports. |
name: ecommerce-marketplace-ops
version: 1.0.0
updated: 2026-02-22
author: InstaClaw
triggers:
keywords: [Shopify, Amazon, eBay, Walmart, orders, inventory, returns, RMA, fulfillment, shipping, marketplace, e-commerce, ecommerce, SKU, listing, pricing, ShipStation]
phrases: ["process this return", "check my orders", "sync inventory", "what sold today", "update my prices", "create a listing", "generate RMA", "daily sales report", "ship this order", "how much did I sell", "low stock alert"]
NOT: [personal shopping, buy this item, add to cart, consumer purchase]
Transforms your InstaClaw agent into a full-time e-commerce operations manager. Connects to Shopify, Amazon, and eBay (user provides their own API credentials — BYOK), syncs inventory across channels, processes returns end-to-end, monitors competitor pricing, and generates unified P&L reports.
What this replaces: $2,000-4,000/month in VAs + SaaS tools for multi-channel sellers ($50k-$5M/year).
All credentials are stored in ~/.openclaw/config/ecommerce.yaml and encrypted at rest via libsodium. Never log credentials. Access restricted to MCP servers only.
# ~/.openclaw/config/ecommerce.yaml
platforms:
shopify:
enabled: true
shop: mystore.myshopify.com
access_token: <encrypted>
amazon:
enabled: true
lwa_client_id: <encrypted>
lwa_client_secret: <encrypted>
refresh_token: <encrypted>
aws_access_key: <encrypted>
aws_secret_key: <encrypted>
seller_id: <encrypted>
marketplace_id: "ATVPDKIKX0DER"
ebay:
enabled: true
app_id: <encrypted>
cert_id: <encrypted>
user_token: <encrypted>
fulfillment:
system: shipstation
api_key: <encrypted>
api_secret: <encrypted>
policies:
return_window_days: 30
auto_approve_threshold: 100
require_human_over: 200
restocking_fee_pct: 0
low_stock_threshold: 10
inventory_buffer_units: 5
max_price_change_pct: 20
| Platform | API Quality | Setup Difficulty | Rate Limits | Cost to User |
|---|---|---|---|---|
| Shopify | 10/10 | EASY (10 min) | 1000 pts/sec GraphQL | FREE |
| Amazon | 8/10 | HARD (30-45 min) | 60-80/hour key endpoints | FREE |
| eBay | 9/10 | MEDIUM (15-20 min) | 5K-50K/day | FREE |
| ShipStation | 8/10 | EASY (5 min) | 40 calls/min | FREE |
| Walmart | Unknown | Unknown | Unknown | FREE (planned) |
| Operation | Shopify | Amazon | eBay | ShipStation |
|---|---|---|---|---|
| Read orders | ✅ | ✅ | ✅ | ✅ |
| Update inventory | ✅ | ✅ | ✅ | ✅ |
| Process returns/RMAs | ✅ | ✅ | ✅ | ✅ |
| Create/edit listings | ✅ | ✅ | ✅ | — |
| Adjust pricing | ✅ | ✅ | ✅ | — |
| Manage fulfillment | ✅ | ✅ | ✅ | ✅ |
| View analytics | ✅ | ✅ | ✅ | ✅ |
| Customer messages | ⚠️ Limited | ✅ | ✅ | — |
| Generate shipping labels | — | — | — | ✅ |
The killer workflow. Fully automated return handling.
Manual process: 10-15 min per return × 10 returns/day = 2+ hours daily. Agent process: Seconds per return, human only inspects item when it arrives.
Return Request → Parse → Fetch Order → Check Eligibility
→ Create RMA → Generate Label → Email Customer → Track Shipment
→ [Item Arrives] → Human Inspects → Agent Processes Refund → Done
fully_autonomous:
- Parse return request
- Fetch order from any platform
- Check eligibility against policy
- Create RMA number
- Generate shipping label
- Email customer with RMA + label
- Track return shipment
- Notify seller when item arrives
human_approval_required:
- Returns outside policy window
- Orders over configured threshold ($200 default)
- Frequent returner flag
- Item condition inspection
- Final refund decision (full/partial/reject)
Shopify (GraphQL):
mutation { returnCreate(input: {
orderId: "gid://shopify/Order/123"
returnLineItems: [{ lineItemId: "...", quantity: 1 }]
}) { return { id, name } } }
mutation { refundCreate(input: {
orderId: "gid://shopify/Order/123"
refundLineItems: [{ lineItemId: "...", quantity: 1 }]
transactions: [{ amount: "29.99", kind: REFUND }]
}) { refund { id } } }
Amazon (SP-API):
returns_api.get_return(return_id)
orders_api.create_refund(order_id=order_id, refund_amount=29.99, refund_reason="CustomerReturn")
eBay (REST):
ebay.sell.return.getReturn(returnId)
ebay.sell.finances.issueRefund({ orderId, refundAmount: { value: "29.99", currency: "USD" } })
Problem: Item sells on Amazon → still shows available on Shopify + eBay → overselling.
Never set platform inventory to exact real count. Always subtract inventory_buffer_units (default: 5) to prevent overselling during sync delays.
max_price_change_pct (20% default)📦 Daily Orders Report — [Date]
SUMMARY
Total Orders: X
Total Revenue: $X.XX
By Platform:
• Shopify: X orders ($X.XX)
• Amazon: X orders ($X.XX)
• eBay: X orders ($X.XX)
NEEDS ATTENTION
X unfulfilled orders
X pending payment
X returns pending
LOW STOCK ALERTS
• SKU-123: 3 remaining (threshold: 10)
• SKU-456: 7 remaining (threshold: 10)
morning_8am:
- Pull overnight orders from all platforms
- Sync inventory across channels
- Check for new return requests
- Generate morning summary → send via Telegram
- Flag low stock items
continuous_monitoring:
every_15_min: Check new orders, sync inventory
every_30_min: Check customer messages across platforms
every_hour: Process return requests
every_2_hours: Monitor competitor pricing
every_6_hours: Full competitive price adjustment run
evening_6pm:
- End-of-day summary (orders, revenue, returns, issues)
- Tomorrow's prep (orders to ship, stock to reorder, returns arriving)
weekly_sunday_8pm:
- Weekly P&L report with charts
- Top/bottom selling products
- Slow-moving inventory recommendations
- Competitor pricing analysis
- Strategic recommendations for next week
| Skill | Integration |
|---|---|
| Email (Skill 8) | Agent emails customers for returns, RMA numbers, refund confirmations |
| Competitive Intel (Skill 10) | Monitor competitor prices, feed into auto-pricing engine |
| Financial Analysis (Skill 7) | Pull sales data for P&L, COGS, margins, platform fee breakdowns |
| Voice & Audio (Skill 3) | Generate audio summary of daily sales for Telegram voice message |
read_orders, write_orders, read_products, write_products, read_inventory, write_inventory, read_fulfillments, write_fulfillmentsecommerce:
enabled: true
platforms:
- shopify
- amazon
- ebay
automations:
returns:
auto_approve_threshold: 100
require_human_approval_over: 200
auto_reject_outside_window: true
inventory:
sync_interval_minutes: 15
low_stock_alert_threshold: 10
buffer_units: 5
pricing:
competitor_monitoring: true
max_price_change_pct: 20
require_human_over_pct: 15
reports:
daily_summary: true
weekly_pnl: true
monthly_analytics: true
notifications:
low_stock: telegram
return_arrived: telegram
order_issues: email
daily_summary: telegram
Phase 1: Core Integration (5 weeks) — MVP
├── Week 1: Install existing MCP servers (Shopify, Amazon, eBay)
├── Week 2: Build onboarding flow + credential storage
├── Week 3-4: RMA workflow automation (killer feature)
└── Week 5: Testing + polish
Phase 2: Fulfillment (2 weeks)
├── Week 6: Build ShipStation MCP server
└── Week 7: Integrate with RMA workflow
Phase 3: Advanced Features (4 weeks)
├── Week 8-9: Multi-channel inventory sync
├── Week 10: Competitive pricing engine
└── Week 11: Analytics/reporting dashboards
Phase 4: Walmart (3-4 weeks, optional)
├── Week 12-14: Build Walmart MCP server + integration
└── Week 15: Testing
TOTAL: 11-15 weeks (3-4 months)
MVP (RMA + orders): 5 weeks
Launch order: Shopify + ShipStation first (low friction) → Amazon (high impact) → eBay (round out top 3) → Walmart (if demand).
Platform Priority Ranking:
| Priority | Platform | Rating | Rationale |
|---|---|---|---|
| 1 | Shopify | 5/5 | Easiest setup, best API, largest market, official MCP exists |
| 2 | Amazon | 5/5 | Largest marketplace, highest revenue/seller, complex setup but worth it |
| 3 | ShipStation | 5/5 | CRITICAL for RMA workflow, most sellers use this, easy API |
| 4 | eBay | 4/5 | Good API, 325-tool MCP exists, easier than Amazon |
| 5 | Walmart | 2/5 | Smallest market, may need custom MCP, lower priority |
Other WMS/Fulfillment Options (Phase 3+):
Recommended approach: ShipStation first (covers 60%+ of multi-channel sellers) → direct carrier APIs → ShipBob/Ordoro on demand.
addressable_market:
shopify_stores: 2,500,000+
amazon_sellers: 1,900,000+
ebay_sellers: 1,300,000+
total: 5,700,000+ (many sell multi-channel)
target_segment:
description: "Multi-channel sellers, $50k-$5M/year, 5-50 orders/day"
estimated_count: 500,000-1,000,000
current_spend:
virtual_assistant: $1,500-3,000/month
ecommerce_saas_tools: $200-500/month
shipstation: $29-159/month
total: $2,000-4,000/month
instaclaw_pricing:
replaces: "VA + multiple SaaS tools"
price_point: $299-999/month
value_prop: "24/7 operation, faster than human, all platforms unified"
revenue_projection:
conservative: "1,000 customers x $299/mo = $299K MRR -> $3.6M ARR"
realistic: "5,000 customers x $499/mo = $2.5M MRR -> $30M ARR"
"Your AI employee for e-commerce operations. InstaClaw agents manage your Shopify, Amazon, and eBay stores 24/7 — processing returns, syncing inventory, monitoring competitors, and generating daily P&L reports. The tedious work that takes you 2+ hours every day? Your agent does it in minutes while you sleep. No more copying RMA numbers between systems. No more overselling because inventory didn't sync. No more missing competitor price drops. Just wake up to a daily summary and approve what needs your attention. Setup takes 15 minutes. First month free."
~/scripts/ecommerce-ops.py — Main operations: orders, inventory, returns, reports~/scripts/ecommerce-setup.sh — Platform credential setup and validation~/.openclaw/skills/ecommerce-marketplace/references/ecommerce-guide.md — API reference