Implement server-level instructions for agent context and usage guidance

[Copilot]

This PR implements server-level instructions for the Fabric RTI MCP server to provide AI agents with comprehensive context on how to effectively use the available tools and services.

Changes Made

Added fabric_rti_mcp/instructions.py

• Comprehensive server instructions (4,128 characters) that provide agents with:
  • Overview of available services (Kusto/Eventhouse and Eventstream)
  • Tool categorization and usage guidance for 19 total tools
  • Common workflows and example usage patterns
  • Security considerations and best practices
  • Configuration guidance and environment variable explanations
  • Getting started steps for new users

Updated fabric_rti_mcp/server.py

• Modified FastMCP instantiation to include the server instructions
• Instructions are now passed to the MCP protocol layer for agent consumption

Added comprehensive tests in tests/test_instructions.py

• 5 test cases covering instruction content, format, and integration
• Validates that instructions contain expected sections and tool references
• Ensures proper FastMCP integration

Example Usage Patterns Covered

The instructions help agents understand when and how to use different services:

### Data Exploration Workflow
1. Start with `kusto_known_services` to see available clusters
2. Use `kusto_list_databases` to find databases of interest  
3. Use `kusto_list_tables` to discover available data
4. Use `kusto_get_table_schema` to understand data structure
5. Run analytical queries with `kusto_query`

### Stream Management Workflow  
1. Use `eventstream_list` to see existing streams
2. Use `eventstream_get_definition` to understand configuration
3. Create new streams with `eventstream_create_simple`

Benefits

• Improved Agent Experience: Agents now receive context about the 19 available tools across both services
• Better Tool Discovery: Clear categorization helps agents choose the right tool for their task
• Security Awareness: Instructions highlight destructive operations and authentication handling
• Reduced Trial-and-Error: Example workflows guide agents through common use cases

The implementation leverages the MCP protocol's server-level instructions feature (introduced in modelcontextprotocol/modelcontextprotocol#148) to provide this context without requiring changes to individual tools.

Fixes #52.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.