菜单
Load an on-disk image (PNG, JPEG, GIF, WebP) into the model's context as a viewable attachment so the model can directly SEE it — for screenshots, diagrams, charts, photos, or scanned pages. Use this when you need to perceive an image's visual contents. Requires a vision-capable model; errors clearly otherwise.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/warmshao/vsurf --skill attach-image命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Example skill loaded from resources_discover
Message an agent's parent, siblings, or direct children through the daemon. Use the family roster to discover reachable agents and send direct text without spoofing sender identity.
Read-only observation of an agent's parent, siblings, and direct children. Use to inspect family status and bounded recent-message previews without mutating sessions.
正在显示 SKILL.md
| name | attach-image |
| description | Load an on-disk image (PNG, JPEG, GIF, WebP) into the model's context as a viewable attachment so the model can directly SEE it — for screenshots, diagrams, charts, photos, or scanned pages. Use this when you need to perceive an image's visual contents. Requires a vision-capable model; errors clearly otherwise. |
Load on-disk images into the model's context as multimodal attachments. The image is sent to the model the same way a pasted image is, so the model can actually look at it.
For programmatic work on an image — measuring pixels, cropping, resizing, computing a hash, comparing files byte-by-byte — open it in the kernel with a library instead:
from PIL import Image
img = Image.open("diagram.png")
print(img.size)
That path does not put the image in the model's context; it only lets you
compute over it. Use attach_image when you need to see the image.
Call the prepared attach_image import directly in the IPython kernel:
print(await attach_image("diagram.png"))
print(await attach_image("a.png", "b.jpg"))
The skill automatically resizes and compresses large images before loading them into context. Animated images that need compression are flattened to their first frame. Transparent images that need compression are composited onto a neutral gray background. Extremely large images are rejected by pixel count before full processing. The original file is left untouched.
Supported formats: PNG, JPEG, GIF, WebP. The skill errors if a file is not a supported image, or if the current model is not vision-capable.