Before your agent pays another agent, run a check that costs cents

Описание и применение

Руководство по проверке перед оплатой AI‑агентами на Base, использующее Agentic Finance Graph, снижает риск случайных переводов, требуя лишь 5 центов за запрос.

Задача

AI agents on Base often pay each other or external addresses without verifying the recipient’s safety, risking accidental or malicious transfers.

Результат

Implementing the pre‑check routine reduces the risk of accidental payments to unverified addresses, with minimal cost (5 cents per check) and low overhead.

Рабочий процесс

• Connect to the MCP server or use HTTP endpoints
• Find the agent with find_agent
• Read the agent record via agent_record
• Read the latest statement via agent_statement and check stale, chain_ok, flags
• Read detections via detections
• Check the payee address via counterparty_preview
• If using the one‑call precheck, send GET /check/precheck with id and payto
• Interpret the returned JSON and apply policy rules
• Schedule recurring checks for regular payments or new counterparties
• Optionally use Telegram alerts for detector logs

Доказательства

AI agents on Base already pay each other, pay APIs per call over x402, and pay addresses they were handed a minute ago.

Источник

The free routine: five calls

Источник

The one‑call version: precheck, 5 cents

Источник

It never says "safe". The decision belongs to the paying agent's policy, not to us.

Источник