PDF Watermark Remover

Use this skill only for an explicit watermark-removal request.

1. Keep the source PDF unchanged and save a sibling copy ending in _no_watermark.pdf.
2. Inspect the PDF first. For separable repeated text layers, run scripts/remove_text_watermark.py; it handles the common /Xi* semi-transparent watermark layer.
3. Render every output page and check that the watermark is gone, while page count, size, main content, and images remain intact.
4. If the watermark is baked into an image or removing it would damage main content, stop and explain instead of covering content blindly.