Feature: Improve repository search in dropdown selector

## Feature Request

Add an easier and more intuitive way to search repositories from the repo dropdown selector.

## Description

The current repository dropdown selector could be enhanced with better search capabilities to help users quickly find and select repositories, especially when dealing with many repos.

## Proposed Implementation

- Add a search/filter input field at the top of the dropdown
- Implement real-time filtering as users type
- Support fuzzy search for partial matches
- Search across repository name, owner, and description
- Add keyboard navigation support (arrow keys, enter to select)
- Show recently accessed repositories at the top
- Add favorites/pinned repositories section

## UI/UX Enhancements

- Clear search button (X) to reset filter
- Highlight matching text in search results
- Show repository metadata (owner, last updated, private/public)
- Group repositories by organization/owner
- Display repository icons/avatars for visual recognition
- Loading state while fetching repositories
- "No results found" message for empty searches

## Benefits

- Faster repository selection
- Better user experience with large repository lists
- Reduced friction in workflow
- Improved accessibility with keyboard support

## Acceptance Criteria

- [ ] Search input field is prominently displayed in dropdown
- [ ] Real-time filtering works smoothly
- [ ] Fuzzy/partial matching is supported
- [ ] Keyboard navigation is fully functional
- [ ] Recently accessed repos are prioritized
- [ ] Search performance is optimized for large lists

🤖 Generated with [Claude Code](https://claude.ai/code)

The issue names the repository dropdown selector but no file, test, or entry point. Start by locating that selector in the Electron app and identify the existing repository-loading and selection behavior. Confirm the intended scope with maintainers, then define completion around the agreed search, navigation, and repository-prioritization criteria.

Written by the indexing model from the issue text.