add-image-vision

To process WhatsApp image attachments for Copilot, this skill automates downloading the images, resizing them with sharp, and encoding them as base64 multimodal content blocks. This pipeline feeds visual data into the agent for interpretation.

Yes, NinjaClaw agents can interpret WhatsApp visual content by merging this skill with the existing WhatsApp skill. It enables image vision by converting attachments into multimodal blocks that Copilot processes for contextual understanding.

Yes, the sharp library is required for image resizing. The skill uses sharp to resize downloaded WhatsApp image attachments before base64 encoding them, ensuring the multimodal content blocks are properly formatted for the Copilot pipeline.

The best way to add image vision to a WhatsApp agent pipeline is merging this skill into your NinjaClaw setup. It handles image download, sharp resizing, and base64 encoding to deliver multimodal content blocks directly to Copilot for richer responses.

WhatsApp image shares fail without base64 encoding because Copilot requires multimodal content blocks, not raw image files. This skill ensures compatibility by downloading, resizing with sharp, and encoding the image data before pipeline integration.