fix(tools): Add proper cleanup for AgentTool to prevent MCP session errors

Merge https://github.com/google/adk-python/pull/3411

## Summary

Fixes AgentTool cleanup to prevent MCP session errors by calling `runner.close()` after sub-agent execution.

## Problem

When using AgentTool with MCP tools, the runner cleanup happened during garbage collection in a different async task context, causing:
```
RuntimeError: Attempted to exit cancel scope in a different task than it was entered in
```

## Solution

- Call `await runner.close()` immediately after sub-agent execution in AgentTool
- This ensures MCP sessions and other resources are cleaned up in the correct async task context
- Updated test mock to include the close() method

## Demo Agents

Added two comprehensive demo agents showing how to use AgentTool with MCP tools:

### mcp_in_agent_tool_remote (SSE mode)
- Uses HTTP/SSE connection to remote MCP server
- Zero-installation setup with `uvx`
- Demonstrates server-side MCP deployment pattern

### mcp_in_agent_tool_stdio (stdio mode)
- Uses subprocess connection with automatic server launch
- Fully automatic setup with `uvx` subdirectory syntax
- Demonstrates embedded MCP deployment pattern

Both demos:
- Use Gemini 2.5 Flash
- Include example prompts for JSON Schema exploration
- Have comprehensive READMEs with architecture diagrams
- Follow ADK agent structure conventions

## Testing

- ✅ All existing unit tests pass
- ✅ Manual testing with both SSE and stdio modes
- ✅ Verified cleanup happens in correct async context
- ✅ No more cancel scope errors with MCP tools

## Related

- Fixes #1112
- Related to #929

Co-authored-by: Wei Sun (Jack) <weisun@google.com>
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/3411 from google:fix/agent-tool-mcp-cleanup 9ae753b5a4
PiperOrigin-RevId: 828651896

contributing/samples/mcp_in_agent_tool_remote/README.md

@@ -0,0 +1,74 @@
+# AgentTool with MCP Demo (SSE Mode)
+
+This demo shows how `AgentTool` works with MCP (Model Context Protocol) toolsets using **SSE mode**.
+
+## SSE vs Stdio Mode
+
+This demo uses **SSE (Server-Sent Events) mode** where the MCP server runs as a separate HTTP server:
+
+- **Remote connection** - Connects to server via HTTP
+- **Separate process** - Server must be started manually
+- **Network communication** - Uses HTTP/SSE for messaging
+
+For the **stdio (subprocess) version**, see [mcp_in_agent_tool_stdio](../mcp_in_agent_tool_stdio/).
+
+## Setup
+
+**Start the MCP simple-tool server in SSE mode** (in a separate terminal):
+
+```bash
+# Run the server using uvx (no installation needed)
+# Port 3000 avoids conflict with adk web (which uses 8000)
+uvx --from 'git+https://github.com/modelcontextprotocol/python-sdk.git#subdirectory=examples/servers/simple-tool' \
+    mcp-simple-tool --transport sse --port 3000
+```
+
+The server should be accessible at `http://localhost:3000/sse`.
+
+## Running the Demo
+
+```bash
+adk web contributing/samples
+```
+
+Then select **mcp_in_agent_tool_remote** from the list and interact with the agent.
+
+## Try These Prompts
+
+This demo uses **Gemini 2.5 Flash** as the model. Try these prompts:
+
+1. **Check available tools:**
+
+   ```
+   What tools do you have access to?
+   ```
+
+2. **Fetch and summarize JSON Schema specification:**
+
+   ```
+   Use the mcp_helper to fetch https://json-schema.org/specification and summarize the key features of JSON Schema
+   ```
+
+## Architecture
+
+```
+main_agent (root_agent)
+  │
+  └── AgentTool wrapping:
+        │
+        └── mcp_helper (sub_agent)
+              │
+              └── McpToolset (SSE connection)
+                    │
+                    └── http://localhost:3000/sse
+                          │
+                          └── MCP simple-tool server
+                                │
+                                └── Website Fetcher Tool
+```
+
+## Related
+
+- **Issue:** [#1112 - Using agent as tool outside of adk web doesn't exit cleanly](https://github.com/google/adk-python/issues/1112)
+- **Related Issue:** [#929 - LiteLLM giving error with OpenAI models and Grafana's MCP server](https://github.com/google/adk-python/issues/929)
+- **Stdio Version:** [mcp_in_agent_tool_stdio](../mcp_in_agent_tool_stdio/) - Uses local subprocess connection

contributing/samples/mcp_in_agent_tool_remote/__init__.py

@@ -0,0 +1,15 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from . import agent

contributing/samples/mcp_in_agent_tool_remote/agent.py

@@ -0,0 +1,70 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from google.adk.agents import Agent
+from google.adk.tools import AgentTool
+from google.adk.tools.mcp_tool import McpToolset
+from google.adk.tools.mcp_tool.mcp_session_manager import SseConnectionParams
+
+# Create MCP toolset
+# This uses the simple-tool MCP server via SSE
+# You need to start the MCP server separately (see README.md)
+mcp_toolset = McpToolset(
+    connection_params=SseConnectionParams(
+        url="http://localhost:3000/sse",
+        timeout=10.0,
+        sse_read_timeout=300.0,
+    )
+)
+
+# Create sub-agent with MCP tools
+# This agent has direct access to MCP tools
+sub_agent = Agent(
+    name="mcp_helper",
+    model="gemini-2.5-flash",
+    description=(
+        "A helpful assistant with access to MCP tools for fetching websites."
+    ),
+    instruction="""You are a helpful assistant with access to MCP tools.
+
+When the user asks for help:
+1. Explain what tools you have available (website fetching)
+2. Use the appropriate tool if needed
+3. Provide clear and helpful responses
+
+You have access to a website fetcher tool via MCP. Use it to fetch and return website content.""",
+    tools=[mcp_toolset],
+)
+
+# Wrap sub-agent as an AgentTool
+# This allows the main agent to delegate tasks to the sub-agent
+# The sub-agent has access to MCP tools for fetching websites
+mcp_agent_tool = AgentTool(agent=sub_agent)
+
+# Create main agent
+# This agent can delegate to the sub-agent via AgentTool
+root_agent = Agent(
+    name="main_agent",
+    model="gemini-2.5-flash",
+    description="Main agent that can delegate to a sub-agent with MCP tools.",
+    instruction="""You are a helpful assistant. You have access to a sub-agent (mcp_helper)
+that has MCP tools for fetching websites.
+
+When the user asks for help:
+- If they need to fetch a website, call the mcp_helper tool
+- Otherwise, respond directly
+
+Always be helpful and explain what you're doing.""",
+    tools=[mcp_agent_tool],
+)

contributing/samples/mcp_in_agent_tool_stdio/README.md

@@ -0,0 +1,74 @@
+# AgentTool with MCP Demo (Stdio Mode)
+
+This demo shows how `AgentTool` works with MCP (Model Context Protocol) toolsets using **stdio mode**.
+
+## Stdio vs SSE Mode
+
+This demo uses **stdio mode** where the MCP server runs as a subprocess:
+
+- **Simpler setup** - No need to start a separate server
+- **Auto-launched** - Server starts automatically when agent runs
+- **Local process** - Uses stdin/stdout for communication
+
+For the **SSE (remote server) version**, see [mcp_in_agent_tool_remote](../mcp_in_agent_tool_remote/).
+
+## Setup
+
+**No installation required!** The MCP server will be launched automatically using `uvx` when you run the agent.
+
+The demo uses `uvx` to fetch and run the MCP simple-tool server directly from the GitHub repository's subdirectory:
+
+```bash
+uvx --from 'git+https://github.com/modelcontextprotocol/python-sdk.git#subdirectory=examples/servers/simple-tool' \
+    mcp-simple-tool
+```
+
+This happens automatically via the stdio connection when the agent starts.
+
+## Running the Demo
+
+```bash
+adk web contributing/samples
+```
+
+Then select **mcp_in_agent_tool_stdio** from the list and interact with the agent.
+
+## Try These Prompts
+
+This demo uses **Gemini 2.5 Flash** as the model. Try these prompts:
+
+1. **Check available tools:**
+
+   ```
+   What tools do you have access to?
+   ```
+
+2. **Fetch and summarize JSON Schema specification:**
+
+   ```
+   Use the mcp_helper to fetch https://json-schema.org/specification and summarize the key features of JSON Schema
+   ```
+
+## Architecture
+
+```
+main_agent (root_agent)
+  │
+  └── AgentTool wrapping:
+        │
+        └── mcp_helper (sub_agent)
+              │
+              └── McpToolset (stdio connection)
+                    │
+                    └── MCP Server (subprocess via uvx)
+                          │
+                          └── uvx --from git+...#subdirectory=... mcp-simple-tool
+                                │
+                                └── Website Fetcher Tool
+```
+
+## Related
+
+- **Issue:** [#1112 - Using agent as tool outside of adk web doesn't exit cleanly](https://github.com/google/adk-python/issues/1112)
+- **Related Issue:** [#929 - LiteLLM giving error with OpenAI models and Grafana's MCP server](https://github.com/google/adk-python/issues/929)
+- **SSE Version:** [mcp_in_agent_tool_remote](../mcp_in_agent_tool_remote/) - Uses remote server connection

contributing/samples/mcp_in_agent_tool_stdio/__init__.py

@@ -0,0 +1,15 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from . import agent

contributing/samples/mcp_in_agent_tool_stdio/agent.py

@@ -0,0 +1,77 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from google.adk.agents import Agent
+from google.adk.tools import AgentTool
+from google.adk.tools.mcp_tool import McpToolset
+from google.adk.tools.mcp_tool.mcp_session_manager import StdioConnectionParams
+from mcp import StdioServerParameters
+
+# Create MCP toolset
+# This uses the simple-tool MCP server via stdio
+# The server will be launched automatically using uvx from the subdirectory
+mcp_toolset = McpToolset(
+    connection_params=StdioConnectionParams(
+        server_params=StdioServerParameters(
+            command="uvx",
+            args=[
+                "--from",
+                "git+https://github.com/modelcontextprotocol/python-sdk.git#subdirectory=examples/servers/simple-tool",
+                "mcp-simple-tool",
+            ],
+        ),
+        timeout=10.0,
+    )
+)
+
+# Create sub-agent with MCP tools
+# This agent has direct access to MCP tools
+sub_agent = Agent(
+    name="mcp_helper",
+    model="gemini-2.5-flash",
+    description=(
+        "A helpful assistant with access to MCP tools for fetching websites."
+    ),
+    instruction="""You are a helpful assistant with access to MCP tools.
+
+When the user asks for help:
+1. Explain what tools you have available (website fetching)
+2. Use the appropriate tool if needed
+3. Provide clear and helpful responses
+
+You have access to a website fetcher tool via MCP. Use it to fetch and return website content.""",
+    tools=[mcp_toolset],
+)
+
+# Wrap sub-agent as an AgentTool
+# This allows the main agent to delegate tasks to the sub-agent
+# The sub-agent has access to MCP tools for fetching websites
+mcp_agent_tool = AgentTool(agent=sub_agent)
+
+# Create main agent
+# This agent can delegate to the sub-agent via AgentTool
+root_agent = Agent(
+    name="main_agent",
+    model="gemini-2.5-flash",
+    description="Main agent that can delegate to a sub-agent with MCP tools.",
+    instruction="""You are a helpful assistant. You have access to a sub-agent (mcp_helper)
+that has MCP tools for fetching websites.
+
+When the user asks for help:
+- If they need to fetch a website, call the mcp_helper tool
+- Otherwise, respond directly
+
+Always be helpful and explain what you're doing.""",
+    tools=[mcp_agent_tool],
+)

src/google/adk/tools/agent_tool.py

@@ -164,6 +164,10 @@ class AgentTool(BaseTool):
         if event.content:
           last_content = event.content
 
+    # Clean up runner resources (especially MCP sessions)
+    # to avoid "Attempted to exit cancel scope in a different task" errors
+    await runner.close()
+
     if not last_content:
       return ''
     merged_text = '\n'.join(p.text for p in last_content.parts if p.text)

tests/unittests/tools/test_agent_tool.py

@@ -131,6 +131,10 @@ async def test_agent_tool_inherits_parent_app_name(monkeypatch):
       )
       return _empty_async_generator()
 
+    async def close(self):
+      """Mock close method."""
+      pass
+
   monkeypatch.setattr('google.adk.runners.Runner', StubRunner)
 
   tool_agent = Agent(