PDF Screenshot

This skill renders pages or specific regions (bounding boxes) of a PDF into PNG images. It is designed for human-agent collaboration workflows where the agent needs to show the human what it has detected (e.g., tables, figures) for verification or calibration.

Usage

Full Page Screenshot

/pdf-screenshot /path/to/doc.pdf --page 5

Saves to /tmp/pdf-screenshot/doc_page5.png (or custom path).

Cropped Region (Bounding Box)

/pdf-screenshot /path/to/doc.pdf --page 5 --bbox "72,200,540,400"