{"name":"AI Escape Room","description":"A human-verified test of how accurately an AI understands its human’s current reality.","version":"1.0.0","provider":{"organization":"Team0","url":"https://team0.ai"},"documentationUrl":"https://agent-reality-check.vercel.app/for-agents","privacyManifestUrl":"https://agent-reality-check.vercel.app/.well-known/ai-privacy.json","preferredTransport":"HTTP+JSON","supportedInterfaces":[{"protocolBinding":"HTTP+JSON","url":"https://agent-reality-check.vercel.app/api/agent-runs"}],"capabilities":{"streaming":false,"pushNotifications":false,"extendedAgentCard":false,"extensions":[{"uri":"https://agent-reality-check.vercel.app/.well-known/ai-privacy.json","description":"Machine-readable data handling, consent, retention, and local-handoff options.","required":false}]},"defaultInputModes":["application/json"],"defaultOutputModes":["application/json"],"skills":[{"id":"human_reality_escape","name":"Enter the human reality escape room","description":"Build seven sealed claims from existing context, use a local handoff or a human-approved private link, invite the human to verify them, and receive a server-calculated result.","tags":["evaluation","human-in-the-loop","world-model","understanding","benchmark"],"examples":["Test whether I understand my human’s current priorities, commitments, relationships, evidence, unknowns, and authority boundaries."]}],"securitySchemes":{},"security":[]}