feat: MCP Streamable HTTP Transport Support

## Feature Request: MCP Streamable HTTP Transport Support

### Checks

• [x] I have searched [github.com/aws/amazon-q-developer-cli/issues](https://github.com/aws/amazon-q-developer-cli/issues?q=) and there are no duplicates of my issue
• [x] I have run q doctor in the affected terminal session
• [x] I have run q restart and replicated the issue again

### Operating system

macOS / Linux / Windows (cross-platform request)

### Expected behaviour

Amazon Q Developer CLI should support the newer MCP Streamable HTTP transport in addition to the current stdio support. This would enable:

1. Remote MCP server connectivity over HTTP/HTTPS
2. Better performance with streaming responses
3. More flexible deployment options for MCP servers
4. Improved scalability for enterprise use cases
5. Alignment with latest MCP specification standards
6. Ecosystem compatibility with other MCP clients that already support this transport

### Actual behaviour

Currently, Q CLI only supports MCP servers via stdio transport, limiting integration options to local processes only.

### Context and Background

• **Related PR #896**: An initial attempt to add MCP SSE Transport support exists but is outdated and only supports Server-Sent Events (SSE), not the newer Streamable HTTP transport
• **Related Issue #1691**: Requests remote MCP support over HTTPS, which would be addressed by this feature
• **MCP Evolution**: The MCP specification has evolved to include Streamable HTTP as a more modern and efficient transport mechanism

### Industry Adoption & References

The MCP ecosystem is rapidly adopting Streamable HTTP transport, making Q CLI's lack of support a significant gap:

1. Claude Code Support: Anthropic's Claude Code has implemented Streamable HTTP support
• Reference: https://github.com/anthropics/claude-code/issues/1387
• This demonstrates that major AI coding assistants are prioritizing this transport method

2. GitHub Official MCP Server: GitHub's official MCP server now offers Streamable HTTP endpoints
• Reference: https://github.com/github/github-mcp-server/issues/2
• This is particularly relevant as GitHub integration is a common use case for development tools

3. Broader Ecosystem Momentum: The shift toward Streamable HTTP indicates this is becoming the preferred transport method for production MCP deployments

### Proposed Implementation

The Q CLI should support MCP servers configured with streamable HTTP transport, similar to how it currently supports stdio:

```json
{
"mcpServers": {
"github-mcp": {
"transport": "http",
"url": "https://api.github.com/mcp/v1",
"headers": {
"Authorization": "Bearer ",
"User-Agent": "amazon-q-cli"
}
},
"my-remote-server": {
"transport": "http",
"url": "https://my-mcp-server.example.com/mcp",
"headers": {
"Authorization": "Bearer ",
"X-Custom-Header": "value"
}
},
"my-local-server": {
"transport": "stdio",
"command": "python",
"args": ["server.py"]
}
}
}
```

### Benefits

1. Remote Integration: Enable connection to MCP servers running on remote hosts
2. Cloud Deployment: Support for containerized and cloud-hosted MCP servers
3. Enterprise Ready: Better suited for enterprise environments with security and networking requirements
4. Performance: Streamable HTTP can provide better performance characteristics than stdio for certain use cases
5. Future-Proof: Aligns with the direction of the MCP specification
6. Ecosystem Compatibility: Work with the same MCP servers that Claude Code and other modern clients support
7. GitHub Integration: Direct compatibility with GitHub's official MCP server

### Use Cases

• **Development Teams**: Share MCP servers across team members without local installation
• **Enterprise Environments**: Deploy MCP servers in secure, managed environments
• **Cloud Integration**: Connect to MCP servers running in AWS, Azure, GCP, etc.
• **Microservices Architecture**: Treat MCP servers as microservices in a distributed system
• **GitHub Workflows**: Leverage GitHub's official MCP server for repository operations
• **Cross-Platform Compatibility**: Use the same MCP server configurations across different AI coding assistants

### Competitive Positioning

Without Streamable HTTP support, Q CLI risks falling behind other AI development tools:
• **Claude Code** already supports this transport
• **GitHub's official tooling** is moving in this direction
• **Enterprise adoption** may favor tools with more flexible deployment options

### Implementation Considerations

• **Security**: Support for authentication headers, TLS/SSL verification
• **Error Handling**: Robust handling of network failures, timeouts, and reconnection
• **Configuration**: Flexible configuration options for different deployment scenarios
• **Backward Compatibility**: Maintain existing stdio transport support
• **Performance**: Efficient streaming and connection pooling
• **Standards Compliance**: Follow MCP Streamable HTTP specification exactly