fix: close daemon health probe socket

[SUJALMU2004]

Summary

• close the IPC socket opened by ensure_daemon() during the warm daemon health probe
• add a regression test for the healthy daemon early-return path

Why

ensure_daemon() probes a live daemon with Target.getTargets before returning. On the healthy path it returned immediately without closing the socket, which can leak descriptors in long-lived agent processes that call ensure_daemon() repeatedly.

Tests

• python -m pytest -q tests\unit\test_admin.py::test_ensure_daemon_closes_probe_socket_when_daemon_is_healthy
• python -m compileall -q src tests
• python -m pytest -q tests\unit\test_admin.py -k "not preserves_snap_bin"

---

Summary by cubic

Fixes a file descriptor leak in the daemon health probe and adds a regression test for the healthy path.

• Bug Fixes
  • Close the probe socket in a finally block and ignore close errors to avoid descriptor leaks on repeated calls.
  • Unit test verifies the healthy path returns early, does not restart the daemon, and closes the socket.

^{Written for commit 4157162. Summary will update on new commits.}

[cubic-dev-ai]

No issues found across 2 files

_{Re-trigger cubic}