feat: Add Copilot support and enable/disable commands for prompts

- Refactor prompts module into a package with handler classes
  - base.py: BasePromptHandler abstract class
  - claude.py, codex.py, gemini.py: Tool-specific handlers
  - copilot.py: GitHub Copilot CLI handler with repo-wide and path-specific instructions
  - manager.py: PromptManager with handler delegation
  - models.py: Prompt dataclass

- Integrate Copilot as --app option in existing commands
  - sync: supports --app copilot with --apply-to for path-specific
  - import-live: imports from .github/copilot-instructions.md
  - show-live: shows Copilot instructions content
  - status: shows Copilot status with path-specific file count
  - unsync: clears Copilot instruction files

- Add enable/disable commands for prompt status control
  - enable: activates prompt, syncs to file, tracks enabled state (user level)
  - disable: deactivates prompt without clearing files

- Remove redundant copilot-sync, copilot-import, copilot-show commands

- Update tests for new functionality

Author: jzhu

.github/copilot-instructions.md

@@ -1,14 +1,9 @@
-# CLAUDE.md — Repository guidelines for AI-assisted edits
+# Copilot Instructions for Testing
 
-This file documents repository-level expectations and instructions intended to guide contributors and AI-assisted editing tools (like Claude Code) when making changes in this project.
+This is a test instruction file for GitHub Copilot CLI custom instructions support.
 
-- Ask for approval before any git commit and push
-- Always run tests before completing all development of new changes
-- Always test the CLI usages
-- After any changes, run the folling to reinstall the project:
-```
-rm -rf dist/*
-./install.sh uninstall
-./install.sh
-cp ~/.config/code-assistant-manager/providers.json.bak ~/.config/code-assistant-manager/providers.json
-```
+## Guidelines
+
+- Write clear, maintainable code
+- Test everything before committing
+- Follow the repository's coding style