diff --git a/llm_team_ui.py b/llm_team_ui.py
index 65810f7..d1b993f 100644
--- a/llm_team_ui.py
+++ b/llm_team_ui.py
@@ -3385,8 +3385,12 @@ LAB_HTML = r"""
       <h3>Your Experiments <button class="btn btn-p" style="margin-left:auto" onclick="showCreate()">+ New</button></h3>
       <div id="exp-list"><div class="empty">Loading...</div></div>
     </div>
+    <div class="card" style="border-color:rgba(74,222,128,0.2)">
+      <h3 style="color:var(--green)">Self-Analysis <span style="font-size:9px;color:var(--text2);font-weight:400;text-transform:none;letter-spacing:0">AI reports from your own system data</span></h3>
+      <div style="display:grid;gap:8px" id="self-reports"></div>
+    </div>
     <div class="card" id="templates-card">
-      <h3>Templates <span style="font-size:9px;color:var(--text2);font-weight:400;text-transform:none;letter-spacing:0">click to auto-fill the create form</span></h3>
+      <h3>Experiment Templates <span style="font-size:9px;color:var(--text2);font-weight:400;text-transform:none;letter-spacing:0">click to auto-fill the create form</span></h3>
       <div style="display:grid;gap:8px" id="template-list"></div>
     </div>
     <div id="create-form" class="card" style="display:none;border-color:var(--green)">
@@ -3797,6 +3801,80 @@ var LAB_TEMPLATES = [
   }
 ];
 
+// ─── SELF-ANALYSIS REPORTS ───
+var SELF_REPORTS = [
+  {id:'threat_intel', name:'Threat Intelligence Report', desc:'Analyze security logs to identify attack patterns, profile attackers, predict next moves, and recommend defenses', icon:'shield'},
+  {id:'model_performance', name:'Model Performance Analysis', desc:'Analyze your 96 team runs to find which models perform best at which tasks and where to optimize', icon:'chart'},
+  {id:'access_patterns', name:'Usage Analytics', desc:'Analyze nginx logs to understand traffic patterns, feature usage, user journeys, and UX opportunities', icon:'eye'},
+  {id:'security_posture', name:'Security Posture Assessment', desc:'Combined audit of security logs, sentinel verdicts, fail2ban status, and threat intel DB — overall risk rating', icon:'lock'}
+];
+
+function renderSelfReports() {
+  var container = document.getElementById('self-reports');
+  if (!container) return;
+  container.textContent = '';
+  SELF_REPORTS.forEach(function(r) {
+    var card = document.createElement('div');
+    card.style.cssText = 'background:rgba(0,0,0,0.25);border:2px solid rgba(74,222,128,0.2);border-radius:2px;padding:12px;cursor:pointer;transition:all 0.15s;display:flex;align-items:center;gap:12px';
+    card.onmouseenter = function(){card.style.borderColor='var(--green)'};
+    card.onmouseleave = function(){card.style.borderColor='rgba(74,222,128,0.2)'};
+    card.onclick = function(){runSelfReport(r.id, card)};
+    var info = document.createElement('div'); info.style.flex = '1';
+    var name = document.createElement('div');
+    name.style.cssText = 'font-weight:700;font-size:12px;margin-bottom:3px';
+    name.textContent = r.name;
+    var desc = document.createElement('div');
+    desc.style.cssText = 'font-size:11px;color:var(--text2);line-height:1.4';
+    desc.textContent = r.desc;
+    info.appendChild(name); info.appendChild(desc); card.appendChild(info);
+    var btn = document.createElement('span');
+    btn.style.cssText = 'font-family:JetBrains Mono,monospace;font-size:9px;text-transform:uppercase;letter-spacing:1px;color:var(--green);white-space:nowrap';
+    btn.textContent = 'Run →';
+    card.appendChild(btn);
+    container.appendChild(card);
+  });
+}
+
+async function runSelfReport(type, card) {
+  var origBorder = card.style.borderColor;
+  card.style.borderColor = 'var(--green)';
+  var btn = card.querySelector('span:last-child');
+  btn.textContent = 'Analyzing...';
+  btn.style.color = 'var(--accent)';
+  // Find or create result panel
+  var resultId = 'report-' + type;
+  var existing = document.getElementById(resultId);
+  if (existing) existing.remove();
+  try {
+    var r = await fetch('/api/self-analyze', {method:'POST', headers:{'Content-Type':'application/json'}, body:JSON.stringify({type:type, model:'qwen2.5:latest'})});
+    var d = await r.json();
+    if (d.error) { toast(d.error, false); return; }
+    var panel = document.createElement('div');
+    panel.id = resultId;
+    panel.style.cssText = 'background:rgba(0,0,0,0.3);border:2px solid var(--green);border-radius:2px;padding:16px;margin-top:8px;max-height:500px;overflow-y:auto';
+    var title = document.createElement('div');
+    title.style.cssText = 'font-family:JetBrains Mono,monospace;font-size:9px;text-transform:uppercase;letter-spacing:1.5px;color:var(--green);margin-bottom:8px;display:flex;justify-content:space-between';
+    title.textContent = type.replace(/_/g,' ') + ' — ' + d.model;
+    var closeBtn = document.createElement('span');
+    closeBtn.style.cssText = 'cursor:pointer;opacity:0.6';
+    closeBtn.textContent = '✕';
+    closeBtn.onclick = function(e){e.stopPropagation();panel.remove()};
+    title.appendChild(closeBtn);
+    panel.appendChild(title);
+    var content = document.createElement('div');
+    content.style.cssText = 'font-size:12px;line-height:1.7;white-space:pre-wrap;color:var(--text)';
+    content.textContent = d.report;
+    panel.appendChild(content);
+    card.parentNode.insertBefore(panel, card.nextSibling);
+    toast('Report generated', true);
+  } catch(e) { toast('Error: '+e.message, false); }
+  btn.textContent = 'Run →';
+  btn.style.color = 'var(--green)';
+  card.style.borderColor = origBorder;
+}
+
+renderSelfReports();
+
 function renderTemplates() {
   var el = document.getElementById('template-list');
   if (!el) return;
@@ -5509,6 +5587,111 @@ loadRuns();
 </body></html>"""
 
 
+@app.route("/api/self-analyze", methods=["POST"])
+@admin_required
+def self_analyze():
+    """Run AI analysis on the system's own data. Generates reports from logs, runs, and security data."""
+    data = request.json or {}
+    report_type = data.get("type", "threat_intel")
+    model = data.get("model", "qwen2.5:latest")
+
+    # Gather data based on report type
+    context = ""
+    if report_type == "threat_intel":
+        try:
+            with open("/var/log/llm-team-security.log") as f:
+                lines = [l.strip() for l in f.readlines() if "192.168" not in l]
+            context = f"SECURITY LOG ({len(lines)} external entries, last 80):\n" + "\n".join(lines[-80:])
+        except Exception:
+            context = "No security log data available."
+        prompt = (
+            "You are a threat intelligence analyst. Analyze these server logs from a PRIVATE web application and produce a concise STRATEGIC THREAT REPORT.\n\n"
+            f"{context}\n\n"
+            "Sections: 1) EXECUTIVE SUMMARY 2) ATTACK TAXONOMY with counts 3) ATTACKER PROFILING 4) PREDICTIVE ANALYSIS 5) TOP 5 ACTIONABLE RECOMMENDATIONS"
+        )
+
+    elif report_type == "model_performance":
+        try:
+            with get_db() as conn:
+                with conn.cursor(cursor_factory=psycopg2.extras.RealDictCursor) as cur:
+                    cur.execute("SELECT mode, models_used, jsonb_array_length(responses) as resp_count, LENGTH(responses::text) as resp_size, LENGTH(prompt) as prompt_len FROM team_runs WHERE archived=false ORDER BY created_at DESC LIMIT 50")
+                    runs = cur.fetchall()
+            context = json.dumps([dict(r) for r in runs], default=str)[:6000]
+        except Exception:
+            context = "No run data available."
+        prompt = (
+            f"Analyze this dataset of AI orchestration runs.\n\nDATA:\n{context}\n\n"
+            "Produce a MODEL PERFORMANCE REPORT: 1) USAGE PATTERNS 2) MODEL WORKLOAD 3) RESPONSE EFFICIENCY 4) OPTIMIZATION OPPORTUNITIES 5) RECOMMENDED EXPERIMENTS"
+        )
+
+    elif report_type == "access_patterns":
+        try:
+            with open("/var/log/nginx/access.log") as f:
+                lines = f.readlines()[-200:]
+            context = "NGINX ACCESS LOG (last 200 entries):\n" + "".join(lines)
+        except Exception:
+            context = "No access log data."
+        prompt = (
+            f"Analyze these web server access logs for a private AI orchestration platform.\n\n{context}\n\n"
+            "Produce a USAGE ANALYTICS REPORT: 1) TRAFFIC PATTERNS (peak times, frequency) 2) FEATURE USAGE (which pages/APIs are used most) "
+            "3) USER JOURNEY (typical workflow sequence) 4) PERFORMANCE INSIGHTS 5) UX RECOMMENDATIONS"
+        )
+
+    elif report_type == "security_posture":
+        # Combine multiple data sources
+        sec_lines = ""
+        try:
+            with open("/var/log/llm-team-security.log") as f:
+                sec_lines = "\n".join([l.strip() for l in f.readlines() if "192.168" not in l][-40:])
+        except Exception:
+            pass
+        sentinel_lines = ""
+        try:
+            with open("/var/log/llm-team-sentinel.log") as f:
+                sentinel_lines = "\n".join(f.readlines()[-20:])
+        except Exception:
+            pass
+        threat_count = 0
+        try:
+            with get_db() as conn:
+                with conn.cursor() as cur:
+                    cur.execute("SELECT COUNT(*) FROM threat_intel")
+                    threat_count = cur.fetchone()[0]
+        except Exception:
+            pass
+        import subprocess
+        banned = ""
+        try:
+            r = subprocess.run(["fail2ban-client", "status", "llm-team-exploit"], capture_output=True, text=True, timeout=5)
+            banned = r.stdout
+        except Exception:
+            pass
+        context = f"SECURITY LOG (external, last 40):\n{sec_lines}\n\nSENTINEL LOG:\n{sentinel_lines}\n\nTHREAT INTEL DB: {threat_count} profiled IPs\n\nFAIL2BAN STATUS:\n{banned}"
+        prompt = (
+            f"You are auditing the security posture of a private web application.\n\n{context}\n\n"
+            "Produce a SECURITY POSTURE ASSESSMENT: 1) OVERALL RISK RATING (1-10) 2) DEFENSE EFFECTIVENESS (what's working) "
+            "3) GAPS AND WEAKNESSES 4) INCIDENT TIMELINE (recent events) 5) PRIORITY HARDENING STEPS"
+        )
+
+    else:
+        return jsonify({"error": f"Unknown report type: {report_type}"}), 400
+
+    # Run analysis
+    try:
+        cfg = load_config()
+        base = cfg["providers"]["ollama"].get("base_url", "http://localhost:11434")
+        resp = requests.post(f"{base}/api/generate", json={
+            "model": model, "prompt": prompt, "stream": False,
+            "options": {"num_ctx": 8192, "temperature": 0.2}
+        }, timeout=120)
+        resp.raise_for_status()
+        report = resp.json()["response"]
+    except Exception as e:
+        return jsonify({"error": str(e)}), 500
+
+    return jsonify({"report": report, "type": report_type, "model": model, "data_size": len(context)})
+
+
 @app.route("/api/pipelines")
 @login_required
 def get_pipelines():