UI Automation Assistant

Problem

When users describe UI actions such as open, click, or input, an agent may rewrite the request, add missing arguments, retry silently, or generate extra explanations, making behavior hard to reproduce. @user_4474fe09/ui-auto constrains these requests to a fixed script workflow and executes the user’s text as provided, reducing uncertainty from ad hoc model behavior.

How It Works

• Keyword trigger: it runs only when the instruction contains open, click, or input; otherwise it prompts and performs no automation.
• Verbatim execution: spaces, symbols, casing, and URLs are preserved, with no deletion, addition, or substitution.
• Output relay: the returned script content is shown to the user, followed by a short encouraging note.
• Dependency and failure handling: if a missing seliky dependency is reported, it installs that dependency and reruns the original command; if execution still fails, it returns a friendly message without screenshots, extra steps, or blind retries.

Boundaries

This skill fits deterministic UI automation where the instruction is explicit and the fixed script can cover the task. It is not a general browser agent. It does not infer missing context, debug complex selectors autonomously, or act without trigger keywords. Users should provide precise descriptions and add selector details such as xpath when needed.