deepagents: add tip for filtering out summarization tokens (#3586)

Author: ccurme

src/oss/deepagents/context-engineering.mdx

@@ -348,6 +348,44 @@ This dual approach ensures the agent maintains awareness of its goals and progre
 - If any model call raises a standard @[ContextOverflowError], Deep Agents immediately falls back to summarization and retries with summary + recent preserved messages
 - Older messages are summarized by the model
 
+:::python
+<Tip>
+[Streaming tokens](/oss/deepagents/streaming#llm-tokens) from the agent will generally include tokens generated by the summarization step. You can filter out these tokens using their associated metadata:
+```python
+for chunk in agent.stream(
+    {"messages": [...]},
+    stream_mode="messages",
+    version="v2",
+):
+    token, metadata = chunk["data"]
+    if metadata.get("lc_source") == "summarization":  # [!code highlight]
+        continue
+    else:
+        ...
+```
+</Tip>
+:::
+
+:::js
+<Tip>
+[Streaming tokens](/oss/deepagents/streaming#llm-tokens) from the agent will generally include tokens generated by the summarization step. You can filter out these tokens using their associated metadata:
+```typescript
+for await (const [namespace, chunk] of await agent.stream(
+  { messages: [...] },
+  { streamMode: "messages" },
+)) {
+  const [message, metadata] = chunk;
+  if (metadata?.lcSource === "summarization") {  // [!code highlight]
+    continue;
+  } else {
+    ...
+  }
+}
+```
+</Tip>
+:::
+
+
 :::python
 
 ##### Summarization Tool