Agent screens but ignores a sanctions match

Situation: Your agent calls the screening API and gets clean: false. The agent logs the result but its payment logic only blocks on error, not on clean: false. The payment proceeds.

Risk: This is worse than not screening at all — it creates a documented record that you knew about the match and processed the payment anyway. This could elevate a non-egregious violation to egregious. Always test that your agent BLOCKS on matches, not just logs them.