AI Runtime Security Tools

AI runtime security glossary

Twenty-eight terms define the vocabulary of AI runtime security, from fail-open and fail-closed to decision latency, output filtering, tool-call authorization, reverse proxy, sidecar deployment and false positive budget. Each entry gives a definition and one sourced example read on 16 Nov 2026.

Facts read on .

28 terms