feat(memory): add LLM-driven memory extraction from conversations

- memory/extractor: async LLM-based memory extraction from chat history
- Parses JSON lines from LLM response as memory entries
- Wire into orchestrator pipeline via PostProcess hook
- Non-blocking: runs in goroutine with 30s timeout
- Automatic dedup via existing keyword overlap logic

backend/services/memory/bridge_llm.go

@@ -0,0 +1,17 @@
+package memory
+
+import (
+	"context"
+
+	"git.yeij.top/AskaEth/Cyrene_v2/services/ai-core/llm"
+	"git.yeij.top/AskaEth/Cyrene_v2/services/ai-core/model"
+)
+
+// LLMAdapterBridge adapts llm.Adapter to LLMClient for memory extraction.
+type LLMAdapterBridge struct {
+	Adapter *llm.Adapter
+}
+
+func (b *LLMAdapterBridge) Chat(ctx context.Context, messages []model.LLMMessage) (*model.LLMResponse, error) {
+	return b.Adapter.Chat(ctx, messages)
+}