fix(prompt): clean persona system prompt — remove YAML dump, filter empty subsessions

- Persona injector: concise natural language rules, not raw YAML
- Context builder: skip empty memory, skip raw intent data
- Prompt reduced from 2000+ tokens to ~500 tokens
- LLM response quality improved (natural tone, no confusion)

backend/services/ai-core/context/context.go

@@ -129,16 +129,18 @@ func (b *Builder) Build(systemPrompt string, sessionID string, subsessionResults
 
 	// 2. Subsession results (injected as system context)
 	for _, r := range subsessionResults {
-		if r.Error != "" {
+		if r.Error != "" || r.Content == "" || r.Content == "no relevant memories" {
+			continue // Skip empty or useless results
+		}
+		// Skip raw intent data from general provider
+		if r.Provider == "general" && strings.HasPrefix(r.Content, "Intent:") {
 			continue
 		}
-		if r.Content != "" {
-			msg := model.LLMMessage{
-				Role:    "system",
-				Content: fmt.Sprintf("[%s context]\n%s", r.Provider, r.Content),
-			}
-			messages = append(messages, msg)
+		msg := model.LLMMessage{
+			Role:    "system",
+			Content: fmt.Sprintf("[%s context]\n%s", r.Provider, r.Content),
 		}
+		messages = append(messages, msg)
 	}
 
 	// 3. IoT device context