test: deflake load-sensitive tests (lmxopcua#769)

- NodeManagerLockTimeBenchmark: wall-clock budget enforced only with OTOPCUA_BENCH=1; default run keeps the state-correctness asserts and logs timings.
- OpcUaClientDriverWire host-status: signal (semaphore) wait that only accepts a non-Running state observed after the server is disposed.
- AbCip PollLatencyComparison: retry once on a fixture connection reset; p95 criterion enforced only with AB_LAT_ENFORCE=1.
- ScriptRunGuardTests: orphan scripts held by a ctx.Released signal instead of a 250 ms wall-clock spin that could finish inside the 60 ms budget under load; stuck-orphan wall-clock rule disabled in the tests that pin the strike counting.
- DependencyMuxReplayTests: split the inside-grace positive control (1 h grace) from the release test (control asserted before unregister), removing the 200 ms grace race.

Claude-Session: https://claude.ai/code/session_01GT67AEHd2Acv1YPwGfXEoV

tests/Core/ZB.MOM.WW.OtOpcUa.Core.Scripting.Tests/FakeScriptContext.cs

@@ -19,6 +19,12 @@ public sealed class FakeScriptContext : ScriptContext
     /// <summary>Gets the log of virtual tag write operations.</summary>
     public List<(string Path, object? Value)> Writes { get; } = [];
 
+    /// <summary>
+    ///     Test gate: a "held" script spins on <c>while (!ctx.Released) { }</c> until the test releases it, so a
+    ///     timeout/orphan scenario depends on a signal rather than a wall-clock spin racing the budget under load.
+    /// </summary>
+    public volatile bool Released;
+
     /// <inheritdoc />
     public override DateTime Now { get; } = new DateTime(2026, 1, 1, 12, 0, 0, DateTimeKind.Utc);