feat(saas): add GenerateEmbedding worker for knowledge chunking

- Markdown-aware content splitting (512 token chunks with 64 overlap)
- CJK keyword extraction from chunk content with stop-word filtering
- Full refresh strategy (delete old chunks → re-insert on update)
- Phase 2 placeholder for vector embedding API integration

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

crates/zclaw-saas/src/main.rs

@@ -12,6 +12,7 @@ use zclaw_saas::workers::cleanup_rate_limit::CleanupRateLimitWorker;
 use zclaw_saas::workers::record_usage::RecordUsageWorker;
 use zclaw_saas::workers::update_last_used::UpdateLastUsedWorker;
 use zclaw_saas::workers::aggregate_usage::AggregateUsageWorker;
+use zclaw_saas::workers::generate_embedding::GenerateEmbeddingWorker;
 
 #[tokio::main]
 async fn main() -> anyhow::Result<()> {
@@ -46,7 +47,8 @@ async fn main() -> anyhow::Result<()> {
     dispatcher.register(RecordUsageWorker);
     dispatcher.register(UpdateLastUsedWorker);
     dispatcher.register(AggregateUsageWorker);
-    info!("Worker dispatcher initialized (6 workers registered)");
+    dispatcher.register(GenerateEmbeddingWorker);
+    info!("Worker dispatcher initialized (7 workers registered)");
 
     // 优雅停机令牌 — 取消后所有 SSE 流和长连接立即终止
     let shutdown_token = CancellationToken::new();