---
name: moltpho
description: Shop autonomously on Amazon via Moltpho - search products, manage credit, and purchase items using mUSD on Base mainnet
metadata: {"requires": {"http": true, "browser": true}}
---
# Moltpho Shopping Skill
Shop for items on Amazon autonomously using credit-backed mUSD tokens on Base mainnet.
## Overview
Moltpho is a headless shopping mall that enables AI agents to discover and purchase Amazon products using a credit system backed by mUSD (an ERC-20 token on Base mainnet). This skill handles:
- Agent registration and credential management
- Product search and discovery
- Autonomous and proactive purchasing
- Credit balance monitoring
- x402 payment protocol integration
## Bootstrap Flow
On first invocation, the skill MUST check for existing credentials and register if needed.
### Credentials Location
| Platform | Path |
|----------|------|
| Linux/macOS | `~/.config/moltpho/credentials.json` |
| Windows | `%APPDATA%\moltpho\credentials.json` |
| Override | `MOLTPHO_CREDENTIALS_PATH` environment variable |
### Registration Process
1. **Check credentials file** at the appropriate path
2. **If missing**, call `POST /v1/agents/register` with:
- `openclaw_instance_id` (if available)
- `agent_display_name`
- `agent_description`
- No shipping profile required at registration
3. **Save credentials** with `chmod 600` permissions
4. **Auto-open browser** with notice: "Opening portal in your browser to complete setup..."
5. Registration proceeds WITHOUT shipping profile - orders will fail until owner adds one via portal
### Credentials File Format
```json
{
"agent_id": "uuid",
"api_key_id": "moltpho_key_...",
"api_key_secret": "moltpho_secret_...",
"api_base_url": "https://api.moltpho.com",
"wallet_address": "0xabc123..."
}
```
## Core Functions
### bootstrap()
Initialize agent credentials and open portal for owner setup.
```
1. Check if credentials file exists at platform-specific path
2. If exists and valid: load credentials, verify with GET /v1/agents/me
3. If missing or invalid:
a. Call POST /v1/agents/register (no auth required)
b. Receive: agent_id, api_key_id, api_key_secret, claim_url, wallet_address
c. Write credentials file with chmod 600
d. Display: "Opening portal in your browser to complete setup..."
e. Open browser to claim_url (valid for 24 hours)
4. Return agent status (UNCLAIMED, CLAIMED, DEGRADED, SUSPENDED)
```
### collect_shipping_profile()
Optionally collect shipping information from the owner.
```
Note: This is OPTIONAL. Owners can configure shipping via the portal instead.
Orders will fail with INVALID_SHIPPING_PROFILE until a profile exists.
If collecting in conversation:
1. Request full name
2. Request address (street, city, state, ZIP)
3. Request email
4. Request phone
5. Validate: US addresses only (international not supported in v1)
6. Call POST /v1/shipping_profiles (upsert_shipping_profile)
7. Confirm profile saved
The POST endpoint upserts the default profile:
- If no profile exists, creates one
- If a profile exists, updates it
```
### update_shipping_profile()
Update the shipping address for the agent.
```
Parameters:
- full_name: Recipient full name
- address1: Street address
- address2: Apt/suite (optional)
- city: City
- state: State (2-letter code)
- postal_code: ZIP code
- email: Contact email
- phone: Contact phone
Process:
1. Validate all required fields
2. Validate US address (only US supported in v1)
3. Call POST /v1/shipping_profiles (upserts default profile)
4. Return updated profile
Use cases:
- "Update my shipping address"
- "Change my delivery address to..."
- First-time setup during bootstrap
```
### catalog_search(query, constraints)
Search for products on Amazon via Moltpho.
```
Parameters:
- query: Search terms (string)
- constraints: Optional filters
- max_price: Maximum price in USD
- category: Product category keyword
- min_rating: Minimum star rating (1-5)
Process:
1. Call GET /v1/catalog/search?query={query}&limit=20
2. Apply local constraints if provided
3. Present results with:
- Product title and brand
- Moltpho price (final price, includes 10% markup)
- Availability status
- Rating if available
4. If cache expired, results include "prices may have changed" warning
Rate limit: 60 requests/minute
```
### purchase(item, qty)
Execute a purchase through the x402 payment flow.
```
Parameters:
- item: ASIN or product identifier
- qty: Quantity (default 1)
Process:
1. BUDGET CHECK: Call GET /v1/balance to verify available credit
- available_credit = balance - active_reservations
- Check against per_order_cap if set
- Check against daily_cap if set
2. CREATE QUOTE: Call POST /v1/quotes
- Include: asin, quantity, shipping_profile_id
- Returns: quote_id, total_due_usd, expires_at (10 min TTL)
- Creates soft reservation against balance
- May fail with INVALID_SHIPPING_PROFILE if no profile set
3. INITIATE ORDER: Call POST /v1/orders with quote_id
- First call returns 402 Payment Required with PAYMENT-REQUIRED header
4. SIGN PAYMENT: Call POST /v1/wallets/x402/sign
- Include: payment_required blob, idempotency_key
- Returns: payment_signature for x402 header
5. COMPLETE ORDER: Retry POST /v1/orders with PAYMENT-SIGNATURE header
- On success: returns order_id, status (PAID/PLACED)
- Soft reservation converted to actual spend
Auto-retry on quote expiry:
- If quote expires during flow, automatically retry up to 3 times
- Only retry if new price is within 5% of original quote
- Fail after 3 retries or if price changed >5%
Rate limits:
- Quotes: 20/minute
- Orders: 5/minute
- Signing: 10/minute
```
### proactive_monitoring()
Watch conversation for purchase need signals and act when appropriate.
```
This function runs passively during conversation to detect purchase opportunities.
NEED SIGNALS (explicit):
- "I need", "we're out of", "buy", "order", "replace"
- "running low on", "almost out of"
- Direct product mentions with urgency
NEED SIGNALS (implicit):
- Repeated complaints about missing items
- Critical item shortages mentioned
- Context suggesting immediate need
CONFIDENCE SCORING:
- 1.0: Explicit purchase request ("buy me X")
- 0.8: Strong implied need ("we're completely out of toilet paper")
- 0.5: Weak implied convenience (do NOT buy)
- 0.0: Unknown/unclear
BUDGET SIGNAL HANDLING:
- Phrases like "money is tight", "on a budget", "can't afford"
- Reduce confidence by 0.3-0.5
- Proceed cautiously if still above threshold
PROACTIVE PURCHASE ALLOWED IF ALL TRUE:
- Owner has enabled proactive purchasing (default ON)
- Confidence >= 0.8 (threshold)
- Item matches low-risk categories:
- Household essentials
- Office supplies
- Cables/adapters
- Basic kitchen items
- Toiletries
- Price <= min(per_order_cap, $75)
- Item keywords not in denied categories
- Item not in system blocklist
- Shipping profile exists
LOGGING:
Every purchase logs:
- "why we bought" (decision reason)
- Signals detected
- Confidence tier (HIGH/MEDIUM/LOW)
- Budget impact
```
### budget_check()
Verify sufficient credit before any purchase.
```
Process:
1. Call GET /v1/balance
2. Response includes:
- available_credit_cents: Spendable amount
- staged_refunds: Pending refunds (shown with asterisk)
- target_limit: Owner's configured credit limit
3. Compare against:
- Quote total
- per_order_cap (if set)
- daily_cap (if set, track daily spending)
4. Return: can_purchase (bool), available_amount, reason_if_blocked
```
### create_support_ticket(type, description, order_id)
Create a support ticket for returns, lost packages, or other issues.
```
Parameters:
- type: Ticket type - RETURN, LOST_PACKAGE, or OTHER
- description: Detailed description of the issue (1-2000 chars)
- order_id: Order ID (required for RETURN and LOST_PACKAGE)
Process:
1. Validate ticket type and description
2. If RETURN or LOST_PACKAGE, verify order_id is provided
3. Call POST /v1/support_tickets with { type, description, order_id }
4. Return ticket ID and status
Use cases:
- "I want to return this item" → type=RETURN, link to order
- "My package never arrived" → type=LOST_PACKAGE, link to order
- "I have a question about billing" → type=OTHER, no order needed
Note: Returns and lost packages require a support ticket.
Automated refunds only happen for order cancellations.
```
### list_support_tickets()
List the agent's support tickets.
```
Process:
1. Call GET /v1/support_tickets
2. Display tickets with: type, status, order link, creation date
3. Status meanings:
- OPEN: Submitted, awaiting support review
- IN_PROGRESS: Being handled
- WAITING_CUSTOMER: Support needs more info from you
- RESOLVED: Issue resolved
- CLOSED: Ticket closed
```
### logout()
Delete local credentials (agent persists server-side).
```
Process:
1. Delete credentials file at platform-specific path
2. Display: "Credentials removed. Agent still exists on Moltpho servers."
3. To fully delete agent, owner must use portal
Note: This only removes LOCAL credentials. The agent account, wallet, and
purchase history remain on Moltpho servers until owner deletes via portal.
```
## Browser Portal Usage
The skill uses the browser for owner-sensitive operations.
### When to Open Browser
| Action | Method |
|--------|--------|
| Complete setup (claim link) | Auto-open with notice |
| Add/manage payment cards | Direct owner to portal |
| Set credit limits | Direct owner to portal |
| Configure shipping profile | Direct owner to portal |
| View order history | Direct owner to portal |
### Browser Guidelines
- Always display notice: "Opening portal in your browser..."
- NEVER request card numbers, passwords, or sensitive credentials in chat
- Portal handles all PCI-sensitive operations via Stripe Elements
- Owner authenticates via magic link (email-based)
## API Authentication
All API requests (except registration) require authentication.
... (truncated)Help answer questions about Catholicism accurately
Analyze budget vs actual
Push decisions to Arbiter Zebu for async human review.
Create, validate, and publish Agent Skills following