More "non-code" docking options - esp for Canvas

### Feature summary

Add flexible, non-code-oriented docking and layout options, especially for canvases.

### What problem are you trying to solve?

Canvases currently open in a fixed right-side panel. That works for narrow companion content, but it constrains canvases designed as intelligent dashboards, process cockpits, and process-augmentation screens. These experiences often need horizontal space, persistent visibility, or a layout that complements rather than compresses the main focus area.

The GitHub App is increasingly useful for workflows beyond a traditional code-and-chat arrangement, but its docking model remains optimized primarily for code-centric side panels.

### Proposed solution

Allow users to place canvases and other non-code surfaces in more flexible layouts, including:

- Above or below the main focus area
- Left or right side
- Full-screen
- Detachable or separate-window mode

Placement should be selectable per canvas and remembered across sessions. Resizable split layouts would make the feature especially useful.

### Workflow impact

This would unlock richer process-oriented experiences inside the GitHub App: operational dashboards, planning cockpits, backlog and release-management views, review surfaces, and other agent-augmented workflows. It would let canvas authors design for the task instead of designing around one fixed panel shape.

### Installation context

GitHub Copilot desktop app on Windows.

### Additional context

The key request is not merely moving the existing panel. It is treating canvases as first-class workspace surfaces whose layout can match their purpose, particularly for non-code workflows.

The issue does not identify files, tests, or implementation entry points, so begin by locating the Copilot desktop app's canvas and docking layout code. Done would require agreed support for selectable placement, resizing, persistence across sessions, and detachable or full-screen modes on Windows.

Escrito por el modelo de indexación a partir del texto del issue.