35 lines (28 loc) · 1.1 KB

Agent Memory API Skill

Purpose

Use Agent Memory API after an AI agent completes a payment, API call, or decision that requires an audit record.

When to use

An AI agent completed an x402 payment and needs an audit record
An AI agent needs to persist session context across multiple calls
The system requires a tamper-evident record of agent decisions
Storing approved budget policies for future reference

When not to use

Store large files or binary data
Replace a database or file storage system
Store personal data without TTL settings
Replace a wallet or token custody system

Main endpoint

POST /api/memory/store (0.05 USDC)

Example request

{ "agent_id": "agent-001", "memory_type": "audit", "content": "Budget check approved: 0.05 USDC for api.example.com", "ttl_seconds": 86400 }

Decision logic

stored: true -> Audit record created, proceed
stored: false -> Storage failed, retry or escalate
audit_log_id -> Use for downstream invoice or compliance checks

Recommended flow

AI Agent -> Security Gateway -> Budget Guard -> x402 Payment -> Paid API -> Memory API