It would enable image-aware automation if the REST API exposed an endpoint that could
return raw image data (or base64) for files referenced in markdown. This would allow external agents
(e.g. Claude integrations) to fetch and describe images without manual uploads or text descriptions.