Enhancement: Clarify Memory Core healthcheck response for hybrid database strategy

The current healthcheck response from the Memory Core can be confusing for agents. It can report a healthy status even when the database process is not managed by the server ("running": false, "managed": false), because the server might be connected to a user-managed, externally running ChromaDB instance.

This leads agents to incorrectly assume the database is down and attempt to start it, causing confusion.

The healthcheck response should be enhanced to make this hybrid startup scenario clearer.

Proposal:

1. Modify the HealthCheckResponse schema in openapi.yaml.
2. Add a new field, perhaps database.process.strategy or similar, with possible values like "managed", "external", or "unknown".
3. Update HealthService.mjs to populate this new field based on whether the connection is to a managed process or an external one.
4. Adjust the human-readable details array to explicitly state the connection type, e.g., "Connected to an externally managed ChromaDB instance."

This will provide a clearer picture to the agent, preventing incorrect assumptions and unnecessary actions.