[FEATURE] Image drag and drop should be implemented to the extention

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

When working in the IDE extension, I frequently need to share visual context with Claude — UI screenshots, design mockups, error dialogs, RTL layout bugs, and diagrams. Right now there's no way to attach an image directly into the extension's input.

Proposed Solution

Let me drag an image file from my OS file manager (or drag/paste from clipboard) directly onto the Claude Code input box in the extension, the same way GitHub's own issue editor supports "Paste, drop, or click to add files."

Specifically:

• Drag a .png/.jpg/.webp from Finder/Explorer onto the input → it attaches inline and is sent to Claude as image content.
• Cmd/Ctrl+V of a clipboard screenshot → same behavior.
• An "attach" affordance (paperclip icon) as a click fallback.
• A small thumbnail preview with a remove (x) button before sending.
• Multiple images per message.

Alternative Solutions

• Currently I screenshot, open the Claude web app, paste there, then copy the answer back into my editor — constant context switching.
• I've tried describing visuals in text, but that loses layout/color/spacing detail that matters for CSS and RTL bugs.
• Some other AI IDE tools (e.g. Cursor) already let you drop images into the chat input, so the pattern is established and expected.

Priority

High - Significant impact on productivity

Feature Category

Developer tools/SDK

Use Case Example

Example scenario:

1. I'm building a bilingual (Arabic RTL + English) React dashboard in VS Code.
2. A component renders correctly in LTR but the spacing/alignment breaks in RTL.
3. I take a screenshot of the broken RTL layout and drag it onto the Claude Code input.
4. Claude sees the actual rendered bug and suggests the exact CSS/logical-property fix, without me leaving the editor or describing the visual in prose.

Additional Context

• GitHub's own issue composer (this very form) supports "Paste, drop, or click to add files" — mirroring that UX in the extension input would feel native.
• The web and desktop Claude apps already accept image input, so the model side is solved; this is an extension-input capability gap.
• Nice-to-have: respect a reasonable per-image size limit and downscale large screenshots client-side before upload.