fix: wrap call_service_tool response in dict to fix Pydantic validation

[brianegge]

Summary

Fixes #29

The Home Assistant API returns a list of affected entity states for service calls (e.g., [] for automation.reload), but call_service_tool was passing this list directly as its return value. Since the function's return type annotation is Dict[str, Any], this caused Pydantic validation errors:

Error executing tool call_service_tool: 1 validation error for call_service_toolOutput
result
  Input should be a valid dictionary [type=dict_type, input_value=[], input_type=list]

Changes

• app/hass.py: Fix call_service return type annotation from Dict[str, Any] to List[Dict[str, Any]] to reflect actual HA API behavior
• app/server.py: Wrap call_service_tool result in a dict with success, domain, service, and affected_entities fields
• tests/test_server.py: Add regression tests for the fix

New Response Format

Before:

[]  # Raw list from HA API - causes validation error

After:

{
    "success": True,
    "domain": "automation",
    "service": "reload",
    "affected_entities": []
}

Test Plan

• Added test_call_service_tool_returns_dict - verifies the tool always returns a dict
• Added test_call_service_tool_with_data - verifies data is passed correctly
• All existing tests pass

🤖 Generated with Claude Code