Add User Model Service and Temporal Recall for MeMind V0.1.

Introduce UMS ingest/snapshot pipeline, Context Planner with multi-source recall, runtime context injection, canonical user mapping, and session snapshot loading on auth/me.

Co-authored-by: Cursor <cursoragent@cursor.com>

direct-chat-service.mjs

@@ -1,4 +1,4 @@
-import crypto from 'node:crypto';
+import { resolveRuntimeContext } from './temporal-recall-service/runtime-context.mjs';
 import { MEMORY_INTERVENTION_LIMIT } from './memory-intervention.mjs';
 import { resolveMemoriesWithLegacyFallback } from './memory-legacy-fallback.mjs';
 import { isMemoryRecallQuestion } from './chat-intent-router.mjs';
@@ -135,10 +135,11 @@ function buildMemorySystemBlock({ memories, routingMemoryContent } = {}) {
   return renderMemoryLines(memories);
 }
 
-function buildModelMessages({ previousMessages, userMessage, memories, routingMemoryContent }) {
+function buildModelMessages({ previousMessages, userMessage, memories, routingMemoryContent, runtimeContextText = '' }) {
   const system = [
     '你是 TKMind H5 聊天助手。',
     '优先直接回答用户问题；不要调用工具。涉及需要执行代码、改文件、生成页面或操作外部系统的任务时，说明当前只能文字讨论，并提示用户开启「深度推理」或用更明确的任务描述。',
+    runtimeContextText,
     buildMemorySystemBlock({ memories, routingMemoryContent }),
   ].filter(Boolean).join('\n\n');
 
@@ -268,6 +269,8 @@ export function createDirectChatService({
   memoryV2 = null,
   conversationMemoryService = null,
   episodicMemoryService = null,
+  pool = null,
+  getUmsPool = null,
   logger = console,
   enabled = envFlag(process.env.MEMIND_DIRECT_CHAT_ENABLED, true),
 } = {}) {
@@ -388,12 +391,23 @@ export function createDirectChatService({
     const memories = routedMemoryContent
       ? []
       : await resolveMemories(userId, activeSessionId, messageText(userMessage));
+    const runtimeContext = await resolveRuntimeContext({
+      pool,
+      getUmsPool,
+      userId,
+      query: messageText(userMessage),
+      sessionId: activeSessionId,
+    }).catch((err) => {
+      logger?.warn?.(`[direct-chat] runtime context skipped: ${err instanceof Error ? err.message : err}`);
+      return { injectionText: '' };
+    });
     const completion = await llmProviderService.createChatCompletion({
       messages: buildModelMessages({
         previousMessages,
         userMessage,
         memories,
         routingMemoryContent: routedMemoryContent,
+        runtimeContextText: runtimeContext.injectionText ?? '',
       }),
     });
     if (!completion?.ok) {