test(workflow): escalation drain loop reassigns then completes (S-14, refs #15)

Failing tests for BeoordelingEscalatieProcessor: acquire the parked
BeoordelingEscaleren jobs, reassign each instance to teamlead, complete the job,
and leave a failed reassignment un-completed for redelivery. PumpOnceAsync is
stubbed to throw so the suite compiles and fails on the new behaviour only.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

services/domain/Big.Infrastructure/BeoordelingEscalatieProcessor.cs

@@ -0,0 +1,19 @@
+using Microsoft.Extensions.Logging;
+
+namespace Big.Infrastructure;
+
+/// <summary>
+/// One poll tick of the beoordeling-escalation worker (S-14, ADR-0015): acquire the parked
+/// <c>BeoordelingEscaleren</c> jobs — the tokens the 14-day boundary timer on <c>Beoordelen</c> spawns
+/// — reassign each instance's still-open <c>Beoordelen</c> task to the teamlead, and complete the job.
+/// A job that fails is logged and left un-completed so Flowable redelivers it (§8.6). Split out from
+/// the hosted pump so the acquire→reassign→complete logic is unit-testable without a running host.
+/// </summary>
+public sealed class BeoordelingEscalatieProcessor(
+    IBeoordelingEscalatieClient client,
+    ILogger<BeoordelingEscalatieProcessor> logger)
+{
+    /// <summary>Acquire and process up to <paramref name="maxJobs"/> escalations. Returns the number acquired.</summary>
+    public Task<int> PumpOnceAsync(int maxJobs, CancellationToken ct = default)
+        => throw new NotImplementedException();
+}