What is OFAC sanctions screening for AI agents?

AI agents that process payments must screen every counterparty, and the legal framework for this is already in place.

TL;DR

TL;DR: OFAC sanctions screening for AI agents means checking a counterparty against the US Treasury SDN List before the agent pays. It is a new compliance frontier, but the rules are not new.

What it is

Screening is a lookup that answers one question: is this counterparty on a sanctions list? For AI agents, it happens as a programmatic step, not a manual review. sanctionsai.dev exposes one HTTP call that screens a counterparty and returns ALLOW or BLOCK in under 100 ms, before any funds move. Clean means the agent proceeds; flagged means the agent stops.

Why it applies to agents

OFAC, the Office of Foreign Assets Control at the US Treasury, enforces strict liability. Intent is not required, and civil penalties start at $356,000 per violation. An autonomous agent that pays a sanctioned wallet has still committed the violation. Integrations like x402, AP2, ACP, and Coinbase AgentKit move money but do not screen, so the screening responsibility falls on whoever builds the agent.

How to add it

Place a screening check before every payment step. The 4-Gate Agent Payment Protocol, SCREEN, SCORE, STOP, STAMP, gives a repeatable sequence. Developers can call the API directly or use the MCP server with pip install sanctions-mcp to expose screening as a tool inside agent workflows.