Support pasting images from clipboard into the conversation

[CaioFML]

Describe the feature or problem you'd like to solve

Support pasting images from clipboard into the conversation

Proposed solution

Allow users to paste images (e.g., screenshots) directly into the CLI input using Ctrl+V / Cmd+V. The image would be sent as a vision-capable attachment to the model, similar to how @
mentions include file contents. The CLI could display a small inline preview or placeholder like [📷 Image pasted from clipboard] before sending. For terminals that don't support inline images, just
showing the placeholder text is fine — the image data would still be sent to the model.

Example prompts or workflows

1. User takes a screenshot of a UI bug → pastes into CLI → asks "What's wrong with this layout?" — the model analyzes the screenshot and suggests CSS/HTML fixes.
2. User pastes a screenshot of an error dialog from a mobile emulator → asks "Fix this error" — no need to manually transcribe the error message.
3. User pastes a design mockup → asks "Implement this component using Tailwind CSS" — the model generates code matching the visual reference.
4. User pastes a screenshot of terminal output that's hard to copy as text (e.g., from a remote machine or a Docker log viewer) → asks "What does this error mean?"

Additional context

Currently, sharing visual context requires either taking a screenshot via Chrome DevTools integration (only works for browser pages) or saving an image to a file and referencing it with
@. Clipboard pasting is a much faster workflow — especially for UI/UX work, debugging visual issues, or when the user wants to show something from outside the terminal (design tools, browser, mobile
emulators). Most AI chat interfaces (ChatGPT, Claude web) already support this, and bringing it to the CLI would significantly reduce friction for visual debugging workflows.