Frontier models miss 57% of threats in agent context

Superagent’s research finds that frontier AI models like Claude 4.6 Opus fail to detect 57% of security threats in agent contexts, missing threats that require reputation data, behavioral history, and trust graphs beyond visible content.