Python: [Bug]: Anthropic connector silently drops ImageContent in user messages and crashes on image-only requests #14531

Describe the bug

In semantic-kernel Python SDK, _format_user_message in semantic_kernel/connectors/ai/anthropic/services/utils.py currently hardcodes:

def _format_user_message(message: ChatMessageContent) -> dict[str, Any]:
    return {
        "role": "user",
        "content": message.content,
    }

This causes two major issues when interacting with multimodal Anthropic Claude models (such as Claude 3.5 Sonnet, Claude 3.7 Sonnet, Claude Opus 4.8, Claude Fable 5.1 / Mythos 5.1):

1. Silent Image Data Loss: If a user message contains ImageContent along with text in message.items, the image items are completely discarded and only the text content is sent to Anthropic.
2. API Crash on Image-Only Messages: If a user message contains only ImageContent without text, message.content is empty/None, causing Anthropic API to return a 400 Bad Request: messages: at least one content block is required error.

Other connectors (Google AI, Vertex AI, AWS Bedrock) iterate over message.items to format ImageContent. Anthropic connector should support this as well.

Steps to Reproduce

from semantic_kernel.contents.chat_message_content import ChatMessageContent
from semantic_kernel.contents.image_content import ImageContent
from semantic_kernel.contents.text_content import TextContent
from semantic_kernel.contents.utils.author_role import AuthorRole
from semantic_kernel.connectors.ai.anthropic.services.utils import _format_user_message

msg = ChatMessageContent(
    role=AuthorRole.USER,
    items=[
        TextContent(text="Analyze this image"),
        ImageContent(data_uri="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg=="),
    ],
)
formatted = _format_user_message(msg)
# formatted['content'] is "Analyze this image" -- image was completely discarded!

Expected Behavior

_format_user_message should inspect message.items and format ImageContent into Anthropic's native {"type": "image", "source": {"type": "base64", ...}} content blocks.