Question: Does AgentCore stdio_client work when pulling docker images

Hi!

I currently have a strands setup where my specialized agents connect to JIRA MCP server via stdio_client using docker command.

Would this setup work with agent core? If not, what is the plan b?

From my tests working locally, when deployed to agent core I have such errors:
MCPClientInitializationError - the client initialization failed`

Follows sample code of the app:

```python
jira_server_params = StdioServerParameters(
command="docker",
args=[
"run",
"-i",
"--rm",
"-e", "CONFLUENCE_URL",
"-e", "CONFLUENCE_PERSONAL_TOKEN",
"-e", "JIRA_URL",
"-e", "JIRA_PERSONAL_TOKEN",
"-e", "JIRA_SSL_VERIFY",
"-e", "CONFLUENCE_SSL_VERIFY",
"ghcr.io/sooperset/mcp-atlassian:latest"
],
env={
"CONFLUENCE_URL": env_vars.get('CONFLUENCE_URL'),
"CONFLUENCE_PERSONAL_TOKEN": env_vars.get('CONFLUENCE_PERSONAL_TOKEN'),
"JIRA_URL": env_vars.get("JIRA_URL"),
"JIRA_PERSONAL_TOKEN": env_vars.get("JIRA_PERSONAL_TOKEN"),
"JIRA_SSL_VERIFY": "false"
}
)

atlassian_mcp_client = MCPClient(lambda: stdio_client(jira_server_params))

@tool
def atlassian_assistant(query: str) -> str:
with atlassian_mcp_client:
tools = atlassian_mcp_client.list_tools_sync()
try:
atlassian_agent = Agent(
model=bedrock_model,
system_prompt="Jira assistant",
tools=tools,
)

processed_response = atlassian_agent(query)
return processed_response
except Exception as e:
return f"Error in Atlassian agent: {str(e)}"

```

Thank you for the help! :)

Start with the deployment behavior around the shown stdio_client configuration and the MCPClientInitializationError. Determine whether AgentCore can pull and run the specified Docker image, then record the supported approach or a concrete alternative for this setup.

Written by the indexing model from the issue text.