[FEATURE]:Screen vision (screenshots) + browser control tools for the agent

Feature hasn't been suggested before.

• I have verified this feature I'm about to request hasn't been suggested before.

Describe the enhancement you want to request

Hi! I'm a user working with an OpenCode agent (Muse Spark) on Windows.
My agent builds GUI programs (tkinter, turtle, pygame) and opens web pages
for me, but it is completely blind:

• It can launch windows (Start-Process) but never sees them. When a program
  crashes on start, it only finds out via log files. It cannot verify layout,
  colors, or that a drawing looks right — it checks code with tests instead
  of looking with eyes.
• It can open URLs but cannot click, scroll, fill forms, or read what is on
  the page. "Click that button for me" is impossible.
  Feature request:

1. A screenshot tool (capture screen / active window) whose output is
   attached as an image the model can actually see.
2. Optionally: basic browser control (go to URL, click, type, read content).
   Use cases: verifying GUI apps the agent just wrote, guiding users step by
   step through websites ("press the blue button"), visual debugging.
   Thank you!