feat: single-instance daemon for shared hyperd across MCP clients

[StefanSteiner]

Summary

• Adds a lightweight daemon that manages a single shared hyperd process per user, so multiple AI clients (Claude Code, Cursor, VS Code Copilot, etc.) share one database engine instead of each spawning their own
• MCP clients auto-discover or auto-spawn the daemon transparently; persistent databases become accessible across all concurrent sessions
• Ephemeral databases are properly cleaned up (DETACH + delete) when sessions end, including on Windows where file locks would otherwise prevent deletion

Architecture

┌──────────────┐  ┌──────────────┐  ┌──────────────┐
│ MCP Client 1 │  │ MCP Client 2 │  │ MCP Client 3 │
└──────┬───────┘  └──────┬───────┘  └──────┬───────┘
       └─────────────────┼─────────────────┘
                         │ libpq (TCP 127.0.0.1)
              ┌──────────┴──────────┐
              │   hyperdb daemon    │
              │  (single instance)  │
              └──────────┬──────────┘
              ┌──────────┴──────────┐
              │       hyperd        │
              └─────────────────────┘

Key design decisions:

• TCP port binding as cross-platform single-instance lock (works identically on Windows/macOS/Linux)
• Discovery file at ~/.hyperdb/daemon.json (overridable via HYPERDB_STATE_DIR)
• Health protocol (PING/HEARTBEAT/STOP/STATUS) for liveness checks and idle tracking
• Idle timeout (30 min default) with debounced heartbeats from active clients
• --no-daemon flag to opt out and use legacy per-client hyperd

New CLI commands

hyperdb-mcp daemon              # Run as daemon (usually auto-spawned)
hyperdb-mcp daemon stop         # Gracefully shut down running daemon
hyperdb-mcp daemon status       # Show running daemon info
hyperdb-mcp --no-daemon         # Legacy mode: private hyperd per session

Test plan

• 26 new tests in daemon_tests.rs covering:
  • DaemonState idle tracking and shutdown flag
  • Discovery file write/read/overwrite/remove/stale-cleanup
  • Health protocol (PING, HEARTBEAT, STOP, STATUS, unknown command, multi-command)
  • Single-instance lock (second bind fails with AddrInUse)
  • Idle timeout triggers shutdown; heartbeats prevent it
  • Full integration: engine connects to shared daemon, two engines share one hyperd
  • Persistent database file survives engine drop
  • Ephemeral database cleaned up on drop (DETACH + delete)
• All existing workspace tests pass unchanged
• cargo clippy and cargo fmt clean
• Manual: start two MCP clients, load data in one, query from other
• Manual: verify --no-daemon spawns private hyperd
• Manual: verify idle timeout shuts down daemon after inactivity