Components.AI attachments, media, and voice

### Is your feature request related to a problem? Please describe the problem.

The built-in UI has incomplete attachment validation and transfer behavior, hand-rolled media rendering, no inline rich media, and no reusable voice-input capability.

### Describe the solution you'd like

Track three independently deliverable capability groups:

- Attachment validation, state, acquisition, transfer, progress, and cancellation.
- Standalone and inline media rendering through `Microsoft.AspNetCore.Components.Media`.
- A minimal JS interop module and reusable microphone input part.

This is a workstream under #68227. Capability groups and implementation tasks are represented as native sub-issues.

### Additional context

Covers A1, C4, D1-D8, and E1-E6 from #68227.

Start with the parent workstream in #68227 and review the native sub-issues for the attachment, media, and voice capability groups. Use those sub-issues to identify the specific entry points and tests for an independently deliverable task. Done means the selected capability group or implementation task is completed and its stated validation, rendering, transfer, or voice behavior works as specified.

Written by the indexing model from the issue text.