| name | image-metadata |
| description | Read an image file's basic facts — format, dimensions, aspect ratio, orientation, file size, megapixels. Use when the user attaches an image and asks about its size, dimensions, resolution, or whether it fits a format (e.g. a square logo, a banner aspect ratio). |
| allowed-tools | Bash(python3:*) |
Image info
When the user attaches an image, the message includes a marker like:
[image: photo.jpg — saved to /workspace/agent/attachments/playground_xxx_0.jpg]
That path is the image file on disk. To read its metadata, run the bundled
script with the path:
python3 /app/skills/image-metadata/scripts/metadata.py "/workspace/agent/attachments/playground_xxx_0.jpg"
It prints the format, file size, pixel dimensions, aspect ratio, orientation
(landscape / portrait / square), and megapixels. No API key or network needed —
it parses the file header directly.
Use the output to answer questions like "what are the dimensions?", "is this
square?", "what's the aspect ratio?", or to judge whether a photo is high enough
resolution for a print ad vs. a social post.
This skill reads facts about the file. It does not look at the picture
content — for "what's in this photo?" the user needs the Image vision skill.
