feat(ai): Automate knowledge base embedding on server startup

The AI knowledge-base server should proactively ensure it's up-to-date when it starts. This improves the agent's initial experience and ensures health checks are accurate.

Acceptance Criteria

• On startup, the knowledge-base MCP server will check for the existence of the dist/ai-knowledge-base.jsonl file.
• If the file does not exist, the server will trigger and await the completion of the full syncDatabase process (create + embed).
• If the file does exist, the server will trigger and await the completion of the embedKnowledgeBase process to sync any changes.
• This process must complete before the initial health check is performed, so the server's status reflects the result of the synchronization.
• The server startup logs should clearly indicate which process is running (sync or embed) and its outcome.