add-image-vision

To enable image vision for WhatsApp agents, this skill downloads image attachments, resizes them using sharp, and encodes them as base64 multimodal content blocks passed directly to the agent for interpretation.

Image processing works by downloading WhatsApp attachments into the workspace, resizing them with sharp for consistent dimensions, and converting them into base64-encoded multimodal content blocks the agent can visually interpret.

Yes, integrating the WhatsApp MCP package is required to enable image vision for group workflows, as it handles message routing and attachment downloads before sharp processes the images into base64 content blocks.

The best way to classify and interpret WhatsApp group images is downloading attachments, resizing them with sharp, and passing base64-encoded multimodal content blocks to the agent for descriptive understanding and response generation.

Agents fail to interpret WhatsApp images without a skill to download attachments, resize them using sharp, and pass them as base64-encoded multimodal content blocks to enable visual understanding.

To set up agent vision, you need the WhatsApp MCP package integration, the sharp image processing library successfully built, and a restart of the Deus service to activate image downloading and base64 encoding.