{"id":"system-agentkit-claim-1","systemId":"system-agentkit","metricLabel":{"EN":"Adversarial guardrail cases","FR":"Adversarial guardrail cases"},"value":"14/14 enforced correctly, reproducible deterministically with zero network or LLM calls","protocol":{"EN":"Adversarial test suite","FR":"Adversarial test suite"},"source":"https://github.com/Yacine-ai-tech/agentkit","limitation":{"EN":"See protocol.","FR":"See protocol."},"verified":true,"permalink":"/api/benchmarks/system-agentkit-claim-1"}