Add intelligent diagnostic agent for model selection errors

[Copilot]

Implements automatic detection and repair of model selection errors caused by outdated CLI versions. When a user specifies an invalid model (e.g., newer model not yet recognized), the agent detects the error, checks for CLI updates, and optionally applies them.

Implementation

DiagnosticAgent Class

• Monitors stderr/stdout with regex patterns for model selection errors
• Checks npm for available CLI updates via npm view @github/copilot version
• Validates config with defaults: enabled: true, autoUpdate: false, autoRepair: true
• Limits repair attempts to 3 to prevent loops

Integration

• Added handleDiagnosticCheck() method to CopilotAutoWrapper for shared error detection
• Invoked on both stdout and stderr data events
• Displays diagnostic summary on process exit

Configuration

// CLI flags
copilot-auto --auto-update "task"
copilot-auto --disable-diagnostics "task"

// Environment
export COPILOT_AUTO_UPDATE=1

// Config file
{
  "diagnostics": {
    "enabled": true,
    "autoUpdate": false,  // Requires explicit opt-in
    "autoRepair": true
  }
}

Safety Considerations

• Auto-update disabled by default, requires explicit --auto-update flag
• Package name validated before executing update command
• Enhanced error messages for common npm failures (not found, network issues)
• All configuration values validated with type checks

Error Detection Patterns

/option\s+'--model[^']*'\s+is\s+invalid/i
/unknown\s+model[:\s]+['"]?([^'".\s]+)['"]?/i
/model\s+['"]?([^'".\s]+)['"]?\s+not\s+found/i

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.