Fix end of files

Author: fayedraza

.github/workflows/dagger-fraud-sim.yml

@@ -26,22 +26,22 @@ jobs:
           GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }}
           DAGGER_CLOUD_TOKEN: ${{ secrets.DAGGER_CLOUD_TOKEN }}
         run: |
-          mkdir -p dashboard
-          dagger call run-fraud-sim --source . --gemini-api-key env:GEMINI_API_KEY --groq-api-key env:GROQ_API_KEY export --path dashboard/summary.json
+          mkdir -p results
+          dagger call run-fraud-sim --source . --gemini-api-key env:GEMINI_API_KEY --groq-api-key env:GROQ_API_KEY export --path results/summary.json
 
       - name: Run AI Agent Analysis
         env:
           GEMINI_API_KEY: ${{ secrets.GEMINI_API_KEY }}
         run: |
           echo "Generating fraud summary insights via AI Agent"
-          dagger call analyze-results --summary dashboard/summary.json > dashboard/insights.txt
+          dagger call analyze-results --summary results/summary.json > results/insights.txt
 
       - name: Create Pull Request
         uses: peter-evans/create-pull-request@v5
         with:
-          commit-message: "Update fraud dashboard with AI insights"
-          title: "Update Fraud Dashboard"
+          commit-message: "Update fraud results with AI insights"
+          title: "Update Fraud Results"
           body: "Automated update of fraud simulation results and AI insights."
-          branch: "fraud-dashboard-update"
+          branch: "fraud-results-update"
           base: "main"
           delete-branch: true

dagger/src/fraud_simulation/process_logs.py

@@ -38,6 +38,8 @@ async def process_logs():
         "Groq": os.environ.get("AI_MODEL_GROQ", "Groq-Llama-3")
     }
 
+    account_stats = {} # user_id -> {total_score, count, alerts}
+
     for log in logs:
         if log.get("event_type") == "signup":
             accounts_created += 1
@@ -95,17 +97,44 @@ async def process_logs():
             "reasons": final_reasons
         }
         results.append(result)
-        total_score += max(primary_ai_score, rule_score)
 
-    avg_score = total_score / len(results) if results else 0
+        # Track per account stats
+        user_id = log["user_id"]
+        score = max(primary_ai_score, rule_score)
+        if user_id not in account_stats:
+            account_stats[user_id] = {"total_score": 0.0, "count": 0, "alerts": 0}
+        account_stats[user_id]["total_score"] += score
+        account_stats[user_id]["count"] += 1
+        if is_anomaly:
+            account_stats[user_id]["alerts"] += 1
+
+        total_score += score
+
+    # Calculate per account metrics
+    account_metrics = []
+    for user_id, stats in account_stats.items():
+        account_metrics.append({
+            "user_id": user_id,
+            "average_score": round(stats["total_score"] / stats["count"], 2) if stats["count"] > 0 else 0,
+            "total_alerts": stats["alerts"]
+        })
+    # Sort by score descending to highlight suspicious accounts
+    account_metrics.sort(key=lambda x: x["average_score"], reverse=True)
 
     output = {
+        "report_type": "Fraud Simulation Summary (Per Iteration)",
         "timestamp": datetime.now().isoformat(),
         "accounts_created": accounts_created,
         "total_logins": len(results),
         "normal_logins": len(results) - suspicious_count,
         "suspicious_logins": suspicious_count,
-        "average_score": round(avg_score, 2),
+        "average_score": round(total_score / len(results) if results else 0, 2),
+        "average_fraud_score_per_account": [
+            {"user_id": m["user_id"], "score": m["average_score"]} for m in account_metrics
+        ],
+        "total_alerts_per_account": [
+            {"user_id": m["user_id"], "alerts": m["total_alerts"]} for m in account_metrics
+        ],
         "details": results
     }
 

dashboard/index.html

@@ -41,4 +41,4 @@
     *   **Rule Set Expansion:** Review and potentially expand the rule-based detection to include other suspicious patterns identified (e.g., logins from new/unusual geographical locations, rapid succession of logins from different IPs).
 4.  **User Education:**
     *   **Security Best Practices:** Periodically educate users on password hygiene, phishing awareness, and the importance of MFA.
-5.  **Broader Anomaly Detection:** Monitor the overall `suspicious_logins` and `average_score` metrics to detect any increasing trends that might indicate a larger attack or a new fraud vector.
+5.  **Broader Anomaly Detection:** Monitor the overall `suspicious_logins` and `average_score` metrics to detect any increasing trends that might indicate a larger attack or a new fraud vector.

dashboard/insights.txt results/insights.txtdashboard/insights.txt renamed to results/insights.txt

@@ -445,4 +445,4 @@
       "reasons": []
     }
   ]
-}
+}