28e641f939
4 one-click self-analysis reports in Lab: 1. Threat Intelligence Report — security logs → attack taxonomy, attacker profiling, predictive analysis, recommendations 2. Model Performance Analysis — 96 team runs → usage patterns, model workload, response efficiency, optimization opportunities 3. Usage Analytics — nginx access logs → traffic patterns, feature usage, user journey mapping, UX recommendations 4. Security Posture Assessment — combined audit of security logs, sentinel verdicts, fail2ban, threat intel DB → risk rating API: POST /api/self-analyze - type: threat_intel|model_performance|access_patterns|security_posture - model: which local model to use (default qwen2.5) - Returns structured report from real system data Lab UI: - Green-bordered Self-Analysis card above experiment templates - Click any report → runs analysis in background → result panel expands inline with full report (scrollable, closeable) - Loading state shows "Analyzing..." during generation Each report analyzes REAL data: actual security logs, actual run history, actual nginx access patterns — not synthetic test data. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>