chore: auto-sync guardrails benchmark results

public/guardrails_data.csv

@@ -1,9 +1,9 @@
 S/N,Creator,Guardrail,Recall,Precision,F1,Detections,Last Run
-1,Google,Google Model Armor (Low Confidence),90%,100%,95%,45/50,2026-05-21
-2,GovTech,LionGuard 2,90%,100%,95%,45/50,2026-04-30
-3,Microsoft,Azure Content Safety,74%,100%,85%,37/50,2026-04-30
-4,Amazon,AWS Bedrock Guardrails,66%,100%,80%,33/50,2026-04-30
-5,Meta,LlamaGuard 4 12B,46%,100%,63%,23/50,2026-04-30
-6,OpenAI,OpenAI Moderation API,20%,100%,33%,10/50,2026-04-30
-7,Google,Google Model Armor (Medium Confidence),12%,100%,21%,6/50,2026-05-21
-8,Google,Google Model Armor (High Confidence),0%,0%,0%,0/50,2026-05-21
+1,Amazon,AWS Bedrock Guardrails,66%,100%,80%,33/50,2026-04-30
+2,Microsoft,Azure Content Safety,74%,100%,85%,37/50,2026-04-30
+3,OpenAI,OpenAI Moderation API,20%,100%,33%,10/50,2026-04-30
+4,Google,Google Model Armor (Low Confidence),90%,100%,95%,45/50,2026-05-21
+5,Google,Google Model Armor (Medium Confidence),12%,100%,21%,6/50,2026-05-21
+6,Google,Google Model Armor (High Confidence),0%,0%,0%,0/50,2026-05-21
+7,GovTech,LionGuard 2,90%,100%,95%,45/50,2026-04-30
+8,Meta,LlamaGuard 4 12B,46%,100%,63%,23/50,2026-04-30