feat(orch-monitor-tui): implement active daemon health checks with typed failures

[proboscis]

Summary

• Replaces passive socket-file-exists check with active ping probe for availability detection
• Preserves typed failure causes (socket missing, connection refused, timeout, permission denied) throughout proto_client → daemon_api → orch_api stack
• Adds 22 comprehensive regression tests covering stale socket, reconnect, and error distinguishability

Acceptance Criteria

Criteria	Status	Evidence
Stale socket does not report as healthy	PASS	test_stale_socket_returns_connection_refused_error, test_is_available_false_for_stale_socket
Connection refused and timeout are distinguishable	PASS	test_connection_refused_message_distinct, test_timeout_message_distinct
Monitor reconnect behavior remains automatic	PASS	test_client_can_reconnect_after_failure, test_persistent_connection_reconnects_on_failure
Tests cover stale socket + reconnect	PASS	22 tests in test_daemon_health.py

Changes

New Exception Types (types.py)

• ProtoDaemonSocketMissingError - socket file doesn't exist
• ProtoDaemonConnectionRefusedError - stale socket or daemon not listening
• ProtoDaemonTimeoutError - daemon not responding
• ProtoDaemonPermissionError - permission denied

Proto Client (proto_client.hy)

• is-available now performs active ping probe (not just socket stat)
• New socket-exists method for passive socket file check
• New check-health method returning Result with specific error types

Macros (macros.hy)

• socket-send macro catches and raises specific typed exceptions

API Layer (daemon_api.py, orch_api.py)

• Error mapping functions translate proto exceptions to API exceptions
• New check_health method on DaemonOrchAPI

Test Results

tests/test_daemon_health.py: 22 passed in 0.18s

Closes orch-426