fix: remove example contamination from neocortex extraction prompt

app/services/neocortex.py

@@ -16,18 +16,25 @@ def extract_and_store_knowledge(text: str, user_id: str = "default_user"):
     api_key = settings.GROQ_API_KEY if settings.GROQ_API_KEY else "dummy_key"
     llm = ChatGroq(model="llama-3.1-8b-instant", api_key=api_key)
     
-    prompt = f"""
-You are the semantic logic center of a brain. Extract major factual entities and their relationships from the text.
+    # Skip extraction for very short inputs (e.g. just a name) — not enough
+    # content to contain meaningful relationships.
+    if len(text.strip().split()) < 4:
+        print(f"Neocortex: Input too short for extraction ({len(text.strip().split())} words), skipping.")
+        return 0
+
+    prompt = f"""You are the semantic logic center of a brain. Extract factual entities and their relationships from the text below.
 Return ONLY a valid JSON array of objects. Each object must have "subject", "relation", and "object" keys.
-Aim for concise, capitalized entity names (e.g., "ALEX", "PYTHON", "BAXTER").
+Use concise, CAPITALIZED entity names.
+
+RULES:
+1. ONLY extract facts that are EXPLICITLY stated in the text. Do NOT invent, guess, or assume anything.
+2. If the text does not contain clear factual relationships, return an empty array: []
+3. Do NOT use any example data. Every triple you return must come directly from the text.
 
 Text: {text}
 
-Example output:
-[
-  {{"subject": "BAXTER", "relation": "IS_A", "object": "DOG"}},
-  {{"subject": "BAXTER", "relation": "LIKES", "object": "TENNIS BALLS"}}
-]
+Return format: [{{"subject": "ENTITY_A", "relation": "RELATION_TYPE", "object": "ENTITY_B"}}]
+If no relationships exist, return: []
 """
     try:
         response = llm.invoke([HumanMessage(content=prompt)])

frontend/src/App.css

@@ -443,6 +443,34 @@
   to { opacity: 1; transform: translateY(0) scale(1); }
 }
 
+body.dark-theme .telemetry-overlay {
+  background: rgba(0, 0, 0, 0.6);
+}
+
+body.dark-theme .telemetry-modal {
+  background: rgba(25, 25, 25, 0.85);
+  border-color: rgba(255, 255, 255, 0.05);
+  color: #fff;
+}
+
+body.dark-theme .modal-header {
+  background: rgba(255, 255, 255, 0.02);
+}
+
+body.dark-theme .modal-header h3 {
+  color: #ccc;
+}
+
+body.dark-theme .modal-header button {
+  background: rgba(255, 255, 255, 0.1);
+  color: #ccc;
+}
+
+body.dark-theme .modal-header button:hover {
+  background: #ff6b35;
+  color: white;
+}
+
 @keyframes feedItemIn {
   from { opacity: 0; transform: translateX(20px); }
   to { opacity: 1; transform: translateX(0); }

frontend/src/components/CognitiveDashboard.css

@@ -154,3 +154,35 @@
   letter-spacing: -0.05em;
   margin-left: auto;
 }
+
+/* ── Dark Mode Overrides ── */
+body.dark-theme .status-card {
+  background: rgba(255, 255, 255, 0.05);
+  border-color: rgba(255, 255, 255, 0.05);
+}
+
+body.dark-theme .status-card h3 {
+  color: #bbb;
+}
+
+body.dark-theme .system-chart {
+  background: rgba(0, 0, 0, 0.2);
+  border-color: rgba(255, 255, 255, 0.05);
+}
+
+body.dark-theme .current-state span {
+  color: #aaa;
+}
+
+body.dark-theme .metric-item {
+  background: rgba(0, 0, 0, 0.2);
+  border-color: rgba(255, 255, 255, 0.05);
+}
+
+body.dark-theme .metric-item strong {
+  color: #fff;
+}
+
+body.dark-theme .metric-info label {
+  color: #aaa;
+}

frontend/vite.config.js

@@ -7,7 +7,7 @@ export default defineConfig({
   server: {
     proxy: {
       '/api': {
-        target: 'http://127.0.0.1:8000',
+        target: 'http://localhost:8000',
         changeOrigin: true,
       }
     }