Add AI triage workflow and LLM documentation

[chagong]

Summary

This PR adds an AI-powered triage workflow to automatically label and comment on new issues, along with LLM documentation that provides guidelines for issue labeling.

Changes

• Added .github/workflows/triage-agent.yml: GitHub Actions workflow that triggers on new issues and uses Azure Functions to automatically analyze and label issues using AI
• Added .github/llms.md: Documentation containing labeling guidelines specific to Java development tools and extensions

Features

• Automatic Issue Triage: When a new issue is opened, the workflow automatically analyzes the issue content and applies appropriate labels
• Manual Trigger: The workflow can also be manually triggered via workflow dispatch for specific issue numbers
• Java Development Focus: The LLM documentation provides context about Java development tools and appropriate labeling categories
• Error Handling: Robust error handling with timeouts and proper HTTP status code validation

Configuration

The workflow is configured specifically for the microsoft/vscode-java-debug repository and includes:

• Proper permissions for reading repository content and writing to issues
• Timeout handling for Azure Function calls
• Comprehensive error reporting

This implementation is based on the successful triage workflow from the vscode-gradle repository, adapted for the Java debugger extension.

Testing

The workflow can be tested by:

1. Opening a new issue (automatic trigger)
2. Using the manual workflow dispatch with an existing issue number

Dependencies

• Requires TRIAGE_FUNCTION_LINK secret to be configured
• Requires TRIAGE_AGENT_TIMEOUT variable to be set
• Uses GitHub token with appropriate permissions