fix(deepagents): propagate subagent lc_agent_name for delegated tasks (#566)

## Summary

Fixes #206.

This updates deepagents subagent delegation so tool executions can
reliably identify the active subagent via
`config.metadata.lc_agent_name` instead of inheriting the parent agent
name. It also adds focused regression tests that validate both compiled
subagents and standard subagent specs follow the same metadata behavior.

## Changes

### `libs/deepagents` subagent metadata propagation

- Updated `createTaskTool` subagent invocation config to explicitly set:
  - `metadata.lc_agent_name = subagent_type`
- existing `configurable.ls_agent_type = "subagent"` behavior remains
unchanged
- Added new regression coverage in `subagent.test.ts`:
  - compiled subagent (`runnable`) path
  - standard subagent spec (`systemPrompt/tools/model`) path
- Tests assert tool-time metadata receives the delegated subagent name
(`worker`), preventing parent-name leakage.

### `libs/deepagents` dependency range update

- Bumped `langsmith` peer dependency range in
`libs/deepagents/package.json` from `>=0.6.0 <1.0.0` to `^0.7.1`.
- Updated `pnpm-lock.yaml` accordingly.

.changeset/subagent-agent-name-metadata.md

@@ -0,0 +1,9 @@
+---
+"deepagents": patch
+---
+
+fix(deepagents): propagate subagent `lc_agent_name` during task delegation
+
+- Ensure `task` tool subagent invocations override `metadata.lc_agent_name` with the selected `subagent_type`.
+- Add regression coverage for both compiled subagents (`runnable`) and standard subagent specs to verify tool-time metadata reflects the active subagent.
+- Update the `langsmith` peer dependency range in `deepagents` to `^0.7.1`.

libs/deepagents/src/middleware/subagent.test.ts

@@ -1,6 +1,7 @@
 import { describe, it, expect, vi } from "vitest";
 import { MemorySaver } from "@langchain/langgraph-checkpoint";
 import { FakeListChatModel } from "@langchain/core/utils/testing";
+import { createAgent, tool } from "langchain";
 import {
   AIMessage,
   BaseMessage,
@@ -9,6 +10,7 @@ import {
   ToolMessage,
 } from "@langchain/core/messages";
 import { RunnableLambda } from "@langchain/core/runnables";
+import { z } from "zod/v4";
 import { CallbackManager } from "@langchain/core/callbacks/manager";
 import { BaseCallbackHandler } from "@langchain/core/callbacks/base";
 import { LangChainTracer } from "@langchain/core/tracers/tracer_langchain";
@@ -834,3 +836,170 @@ describe("ls_agent_type tracing metadata", () => {
     expect(subagentRuns.length).toBeGreaterThan(0);
   });
 });
+
+describe("lc_agent_name propagation for subagents", () => {
+  it("should pass subagent name for compiled subagents", async () => {
+    let capturedSubagentAgentName: string | undefined;
+
+    const identifyCaller = tool(
+      (_input, config) => {
+        capturedSubagentAgentName = config.metadata?.lc_agent_name as
+          | string
+          | undefined;
+        return "captured";
+      },
+      {
+        name: "identify_caller",
+        description: "Capture lc_agent_name from metadata",
+        schema: z.object({}),
+      },
+    );
+
+    const compiledSubagentModel = new FakeListChatModel({
+      responses: [
+        new AIMessage({
+          content: "",
+          tool_calls: [
+            {
+              id: "compiled_tool_call",
+              name: "identify_caller",
+              args: {},
+            },
+          ],
+        }) as unknown as string,
+        "Subagent done",
+      ],
+    });
+
+    const compiledSubagent = createAgent({
+      model: compiledSubagentModel,
+      systemPrompt:
+        "Use identify_caller to capture who invoked this subagent, then finish.",
+      tools: [identifyCaller],
+      name: "compiled-worker-inner",
+    });
+
+    const parentModel = new FakeListChatModel({
+      responses: [
+        new AIMessage({
+          content: "",
+          tool_calls: [
+            {
+              id: "task_call_compiled",
+              name: "task",
+              args: {
+                description: "Do work",
+                subagent_type: "worker",
+              },
+            },
+          ],
+        }) as unknown as string,
+        "Done",
+      ],
+    });
+
+    const agent = createDeepAgent({
+      model: parentModel,
+      name: "main-agent",
+      subagents: [
+        {
+          name: "worker",
+          description: "A worker agent",
+          runnable: compiledSubagent,
+        },
+      ],
+    });
+
+    await agent.invoke(
+      { messages: [new HumanMessage("Test")] },
+      {
+        configurable: {
+          thread_id: `test-lc-agent-name-compiled-${Date.now()}`,
+        },
+        recursionLimit: 50,
+      },
+    );
+
+    expect(capturedSubagentAgentName).toBe("worker");
+  });
+
+  it("should pass subagent name for standard subagent specs", async () => {
+    let capturedSubagentAgentName: string | undefined;
+
+    const identifyCaller = tool(
+      (_input, config) => {
+        capturedSubagentAgentName = config.metadata?.lc_agent_name as
+          | string
+          | undefined;
+        return "captured";
+      },
+      {
+        name: "identify_caller",
+        description: "Capture lc_agent_name from metadata",
+        schema: z.object({}),
+      },
+    );
+
+    const standardSubagentModel = new FakeListChatModel({
+      responses: [
+        new AIMessage({
+          content: "",
+          tool_calls: [
+            {
+              id: "standard_tool_call",
+              name: "identify_caller",
+              args: {},
+            },
+          ],
+        }) as unknown as string,
+        "Subagent done",
+      ],
+    });
+
+    const parentModel = new FakeListChatModel({
+      responses: [
+        new AIMessage({
+          content: "",
+          tool_calls: [
+            {
+              id: "task_call_standard",
+              name: "task",
+              args: {
+                description: "Do work",
+                subagent_type: "worker",
+              },
+            },
+          ],
+        }) as unknown as string,
+        "Done",
+      ],
+    });
+
+    const agent = createDeepAgent({
+      model: parentModel,
+      name: "main-agent",
+      subagents: [
+        {
+          name: "worker",
+          description: "A worker agent",
+          systemPrompt:
+            "Use identify_caller to capture who invoked this subagent, then finish.",
+          tools: [identifyCaller],
+          model: standardSubagentModel,
+        },
+      ],
+    });
+
+    await agent.invoke(
+      { messages: [new HumanMessage("Test")] },
+      {
+        configurable: {
+          thread_id: `test-lc-agent-name-standard-${Date.now()}`,
+        },
+        recursionLimit: 50,
+      },
+    );
+
+    expect(capturedSubagentAgentName).toBe("worker");
+  });
+});

libs/deepagents/src/middleware/subagents.ts

@@ -627,6 +627,10 @@ function createTaskTool(options: {
       // Invoke the subagent with ls_agent_type metadata for LangSmith tracing
       const subagentConfig = {
         ...config,
+        metadata: {
+          ...config.metadata,
+          lc_agent_name: subagent_type,
+        },
         configurable: {
           ...config.configurable,
           ls_agent_type: "subagent",