Menu
Analyze images from local files, URLs, or base64 data. For user-uploaded images, the path is provided in the [image attached] annotation.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Platform context for tvOS. No invoke_platform functions are available on Apple TV.
Platform functions available on watchOS. Use invoke_platform to call native watchOS capabilities like local notifications.
Memory system with daily logs and keyword search. Use to save facts, search past events, or recall user preferences and project context across conversations.
Full browser automation via Chrome DevTools Protocol. Navigate pages, interact with UI elements, take screenshots, extract content, manage tabs, emulate devices, handle cookies/storage, monitor network, and control browser state.
Make HTTP requests (GET, POST, PUT, PATCH, DELETE, HEAD, OPTIONS) to APIs and web services. Use when the user needs to call a REST API, fetch data from an endpoint, send form data, upload files, download files, or test HTTP services.
Create and edit images locally with no AI (programmatic operations). Use when the user wants to create a new image (blank, gradient, solid color), resize an image, draw rectangles or shapes on an image, add a watermark, paste a logo, overlay one image on another, or do any Pillow/ImageMagick-style image operations. Do not use for text-to-image generation — use generate_image (AI) instead.
| name | vision |
| description | Analyze images from local files, URLs, or base64 data. For user-uploaded images, the path is provided in the [image attached] annotation. |
Analyze images using the vision tool. When a user uploads an image, you receive an [image attached: path] annotation — use that path as the path parameter.
| Parameter | Type | Required | Description |
|---|---|---|---|
path | string | one of path/url/base64 | Path to a local image file (relative to project root, or absolute for temp files) |
url | string | one of path/url/base64 | URL of a remote image to fetch and analyze |
base64 | string | one of path/url/base64 | Base64-encoded image data (with or without data URI prefix) |
question | string | no | Specific question about the image. If omitted, provides a general description |
mime_type | string | no | Override MIME type (auto-detected from file extension or URL) |
The vision tool loads an image from any source, resizes it to an LLM-friendly preview (max 1024px wide, JPEG), and passes it to the model as a visual content block. The model then "sees" the image and can describe, analyze, or answer questions about it — no external vision API required.
vision path="public/media/photo.jpg"
vision path="public/screenshots/page.png" question="What error message is shown?"
vision url="https://example.com/chart.png" question="What trends does this chart show?"
vision path="public/documents/scan.png" question="Extract all text from this image"
browser action="screenshot" output_path="public/screenshots/page.png"
vision path="public/screenshots/page.png" question="What is displayed on this page?"
vision base64="data:image/png;base64,iVBOR..." question="What is this?"
JPEG, PNG, GIF, WebP, SVG, BMP, ICO, TIFF, AVIF.
Max image size: 20MB.
path, url, or base64question to focus the analysis on specific aspects