How to test OFAC screening integration

A test plan for proving a screening integration blocks flagged counterparties and allows clean ones.

TL;DR

TL;DR: Test the integration with blocked and clean fixtures, cover the fail-closed path, and re-run the suite on a schedule against live data.

Fixture tests

Build a fixture set of known blocked and clean counterparties. Assert that a blocked SDN name or listed wallet returns BLOCK from sanctions_check and that a clean counterparty returns ALLOW. These fixtures make regressions obvious, since a single wrong result means the gate broke.

Fail-closed and edge tests

Test what happens when the screening service is unreachable. The integration should stop the payment, not allow it. Add edge cases too: name variants and near-misses that should still match, plus the 50 Percent Rule, where an entity owned 50 percent or more by a blocked person is itself blocked.

Scheduled re-validation

Sanctions lists change, so run the suite on a schedule and after every deployment. The sanctionsai.dev data set is synced hourly across 16 jurisdictions, so re-running against current data confirms the gate still reflects the latest SDN List. Record results as proof the screen was active and correct.