deepseek-v4-vision-ocr

Use this skill whenever the user sends an image and you cannot see it directly (e.g., you are running on a non-multimodal model like deepseek-v4-pro). This skill runs Tesseract OCR locally to extract text from images. Use when: the user pastes/sends an image, asks to "read" or "recognize" an image, or asks what's in a picture/screenshot. Also use when the user asks for OCR or text extraction from any image file.