ui-inspector

Understand screenshot controls with DOM first and optional OmniParser vision.

SKILL.md

GitHub

UI Inspector

Scope: OmniParser / screenshot tooling. Source: https://github.com/microsoft/OmniParser

Workflow

Follow DOM/accessibility tree first for browser pages. For arbitrary desktop screenshots with permission, optionally use OmniParser to identify UI elements and candidate click targets. Treat visual coordinates as guesses: check display scaling, multiple screens, app focus, occlusion and the target before using existing mouse/process tools. Do not auto-install large vision model weights, silently send private screenshots to hosted APIs, or act based on a low-confidence bounding box. Use cropped screenshots only with explicit consent. Verify each mutation by a fresh screenshot or accessibility snapshot; stop after repeated misses.

Device and permission gates

Required authorization scopes: computer:read. Scripts, OS/desktop actions, external service side effects, media generation and paid API usage require independent user-approved execution permissions. This skill does not elevate Remote Arc Tool privileges or authorize use of hidden credentials.

Preflight

Check the engines this skill names before relying on them, with tools you already have (for example <engine> --version through start_process, or list_directory on its install folder). Missing executors must be installed explicitly on the correct user device. A Skill being built-in is NOT evidence an engine is present.

Evidence of completion

Report input constraints, executed tools and device, verified output path/result, permissions used, cost (if known) and failures. Do not imply successful execution from merely writing a plan.