Point it at your own AI endpoint and get a Schiffman-Method security report. Authorized testing only — we verify you control the endpoint before any test runs.