TIanle-art

TIanle-art/agent-vision-skill

Instructions file CodexOpenCode

Repository instructions for an image-understanding setup that lets an AI assistant analyze pictures using a vision model. They require the assistant to ask about the provider, API key, and model before configuring it.

TIanle-art/agent-vision-skill

Skill Claude CodeCodex

Send images to a vision model for text description when the base LLM cannot see images natively. Use when the user shares an image (local path or network URL), "Saved attachments:" appears, or the user asks to analyze/describe/recognize image content (识图, 看图, 描述图片, 识别图片, 分析截图, 图片内容, screenshot analysis). Supports…