Skill Claude CodeCodex
Low-hallucination image analysis powered by a local/cloud multimodal model via the vision-mcp server. Use when the user wants to analyze, describe, OCR, or detect elements in an image AND accuracy matters more than fluency — e.g. "看看这张图里有什么"、"截图分析"、"提取图片文字"、"识别图中物体". Activates for any image-understanding request where…
0 2mo ago A 110 tokens