test(cli): fix order-dependent test_resume_quiet_stderr flake at the source

test_session_not_found_goes_to_stdout_in_full_mode passes in isolation but
fails in a full tests/cli run. Two independent leaks from the same neighbor
test conspire:

1. test_cli_init.py's _make_cli() reloads cli.py while prompt_toolkit is
   stubbed with MagicMocks and never reloads it back, so sys.modules['cli']
   is left with a mock _pt_print/_PT_ANSI and cli._cprint silently no-ops
   for every later test. Fixed by reloading cli once more with the real
   modules visible (try/finally).

2. prompt_toolkit's print_formatted_text caches its Output on the
   process-global default AppSession the first time it renders without an
   explicit output=. Under capsys (which swaps sys.stdout per test), the
   first CLI test to emit through _cprint locks that cache onto its own
   captured stdout, so later capsys tests read an empty buffer. Fixed with
   an autouse fixture in a new tests/cli/conftest.py that resets the cached
   output around each test.

Neither change touches production code or the flaky test's own assertion.

Related to #59358 (which addresses the same flaky test by mocking _cprint in
the assertion instead; this fixes the two underlying leaks at the source and
does not modify test_resume_quiet_stderr.py).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

tests/cli/test_cli_init.py

@@ -44,13 +44,28 @@ def _make_cli(env_overrides=None, config_overrides=None, **kwargs):
         "prompt_toolkit.formatted_text": MagicMock(),
         "prompt_toolkit.auto_suggest": MagicMock(),
     }
-    with patch.dict(sys.modules, prompt_toolkit_stubs), \
-         patch.dict("os.environ", clean_env, clear=False):
-        import cli as _cli_mod
-        _cli_mod = importlib.reload(_cli_mod)
-        with patch.object(_cli_mod, "get_tool_definitions", return_value=[]), \
-             patch.dict(_cli_mod.__dict__, {"CLI_CONFIG": _clean_config}):
-            return _cli_mod.HermesCLI(**kwargs)
+    try:
+        with patch.dict(sys.modules, prompt_toolkit_stubs), \
+             patch.dict("os.environ", clean_env, clear=False):
+            import cli as _cli_mod
+            _cli_mod = importlib.reload(_cli_mod)
+            with patch.object(_cli_mod, "get_tool_definitions", return_value=[]), \
+                 patch.dict(_cli_mod.__dict__, {"CLI_CONFIG": _clean_config}):
+                return _cli_mod.HermesCLI(**kwargs)
+    finally:
+        # The reload above re-executed cli.py while prompt_toolkit was stubbed
+        # with MagicMocks, permanently rebinding cli's module globals
+        # (``_pt_print``, ``_PT_ANSI``, …) to those mocks. ``patch.dict``
+        # restores ``sys.modules`` on exit, but NOT the names the reloaded
+        # module already bound — so ``sys.modules["cli"]`` is left with a
+        # mock ``_pt_print``, and ``cli._cprint`` then silently no-ops for
+        # every later test (one half of the order-dependent
+        # ``test_resume_quiet_stderr`` full-suite failure; the other half is
+        # the prompt_toolkit output cache reset in this dir's conftest).
+        # Reload once more with the real modules visible so cli's globals
+        # rebind cleanly.
+        import cli as _cli_restore
+        importlib.reload(_cli_restore)
 
 
 class TestMaxTurnsResolution: