diff --git a/‎.changeset/22590-run-summary-acted.md‎
Lines changed: 9 additions & 0 deletions b/‎.changeset/22590-run-summary-acted.md‎
Lines changed: 9 additions & 0 deletions
diff --git a/‎packages/services/service-automation/src/sys-automation-run-first-filter-highlight.test.ts‎
Lines changed: 42 additions & 0 deletions b/‎packages/services/service-automation/src/sys-automation-run-first-filter-highlight.test.ts‎
Lines changed: 42 additions & 0 deletions
diff --git a/‎packages/services/service-automation/src/sys-automation-run.object.ts‎
Lines changed: 12 additions & 5 deletions b/‎packages/services/service-automation/src/sys-automation-run.object.ts‎
Lines changed: 12 additions & 5 deletions
@@ -0,0 +1,9 @@
+---
+'@objectstack/service-automation': patch
+---
+
+fix(service-automation): the `sys_automation_run` row's highlight set now shows `unmeasured_count` beside `selected_count` and `acted_count`. That puts all three operands of the broken-sweep first filter (`selected_count > 0 AND acted_count = 0 AND unmeasured_count = 0`) on the run row itself. The field's description now also lists a notification queued for delivery as an uncountable effect.
+
+Why: on a stack composed as `serve` composes it, reliable delivery is on, so a `notify` step's in-app message is enqueued. The outbox dispatcher writes it after the run settles, and the node reports `unmeasuredEffect`. A notify sweep that did its job therefore reads `acted 0, unmeasured 1`. With only `selected` and `acted` on the row, it read the same as a sweep with nothing to do. No counter changes: an enqueued delivery stays uncountable, never a fabricated `acted`.
+
+For operators: read the triple (`selected`, `acted`, `unmeasured`), never `acted` alone. `acted 0` with `unmeasured` above 0 means "cannot tell", not "did nothing".
@@ -0,0 +1,42 @@
+// Copyright (c) 2026 ObjectStack. Licensed under the Apache-2.0 license.
+
+import { describe, it, expect } from 'vitest';
+import { SysAutomationRun } from './sys-automation-run.object.js';
+
+/**
+ * `sys_automation_run` — the run row shows the WHOLE broken-sweep first filter.
+ *
+ * The filter is `selected_count > 0 AND acted_count = 0 AND unmeasured_count = 0`
+ * (#4354, qualified as a filter and not a verdict by #12685). `highlightFields`
+ * is how the object puts it on the run row itself, and the third operand is the
+ * one that matters for reading the other two: `acted_count = 0` beside a
+ * non-zero `unmeasured_count` means "cannot tell", not "did nothing".
+ *
+ * Measured on a stack composed as `serve` composes it (#22590): a flow whose
+ * only effect is a `notify` step answers `selected 1, acted 0, unmeasured 1`,
+ * because with reliable delivery on the in-app message is enqueued and written
+ * by the outbox dispatcher after the run settles. With `unmeasured_count` off
+ * the highlight set, that delivering sweep read `selected 1, acted 0` on the
+ * run row: the same row as a sweep with nothing to do.
+ *
+ * ⛔ Do not "fix" a failure here by narrowing the list below. An operand leaves
+ * the highlight set only together with the filter expression it belongs to.
+ */
+const FIRST_FILTER_OPERANDS = ['selected_count', 'acted_count', 'unmeasured_count'] as const;
+
+describe('sys_automation_run — the highlight set carries every operand of the first filter', () => {
+  const fields = SysAutomationRun.fields as Record<string, Record<string, unknown>>;
+
+  it('each operand is a declared number column (so the containment below is not a typo match)', () => {
+    for (const name of FIRST_FILTER_OPERANDS) {
+      expect(fields[name], `${name} is expected to be a column`).toBeDefined();
+      expect(fields[name].type).toBe('number');
+    }
+  });
+
+  it('highlightFields contains selected_count, acted_count AND the qualifier unmeasured_count', () => {
+    const highlight = SysAutomationRun.highlightFields ?? [];
+    const missing = FIRST_FILTER_OPERANDS.filter((name) => !highlight.includes(name));
+    expect(missing, `first-filter operands missing from highlightFields: ${missing.join(', ')}`).toEqual([]);
+  });
+});
@@ -77,10 +77,17 @@ export const SysAutomationRun = ObjectSchema.create({
   displayNameField: 'display_title',
   nameField: 'display_title', // [ADR-0079] canonical primary-title pointer (mirrors deprecated displayNameField)
   titleFormat: '{flow_name} · {node_id}',
-  // `selected_count`/`acted_count` sit in the highlight set on purpose (#4354):
-  // "selected 30, acted 0" has to be visible on the run row itself, not one
-  // drill-down away — a signal you must click to find is a signal nobody sees.
-  highlightFields: ['flow_name', 'node_id', 'status', 'selected_count', 'acted_count', 'correlation', 'started_at', 'updated_at'],
+  // The three operands of the broken-sweep first filter —
+  // `selected_count > 0 AND acted_count = 0 AND unmeasured_count = 0` — sit in
+  // the highlight set on purpose (#4354): "selected 30, acted 0" has to be
+  // visible on the run row itself, not one drill-down away — a signal you must
+  // click to find is a signal nobody sees. The qualifier is not optional there:
+  // `acted_count = 0` beside a non-zero `unmeasured_count` means "cannot tell",
+  // not "did nothing". On a served stack a notify step's delivery is enqueued
+  // (reliable delivery on), so a notify sweep that did its job reads
+  // `acted 0, unmeasured 1` — and with the qualifier off the row it read as a
+  // sweep with nothing to do (#22590).
+  highlightFields: ['flow_name', 'node_id', 'status', 'selected_count', 'acted_count', 'unmeasured_count', 'correlation', 'started_at', 'updated_at'],
 
   fields: {
     id: Field.text({ label: 'Run ID', required: true, readonly: true, group: 'System' }),
@@ -459,7 +466,7 @@ export const SysAutomationRun = ObjectSchema.create({
     unmeasured_count: Field.number({
       label: 'Uncountable Effects',
       required: false,
-      description: 'Executions that reached something the platform cannot count (a `connector_action`, a mutating `http` call whose response was lost). The qualifier `acted_count` needs to be trusted: the broken-sweep filter is `selected_count > 0 AND acted_count = 0 AND unmeasured_count = 0` (a filter, not a verdict — see `acted_count`), because a run with uncountable effects has an INCOMPLETE acted count, not a zero one. Null on rows written before this was tracked.',
+      description: 'Executions that reached something the platform cannot count (a `connector_action`, a mutating `http` call whose response was lost, a notification queued for delivery). The qualifier `acted_count` needs to be trusted: the broken-sweep filter is `selected_count > 0 AND acted_count = 0 AND unmeasured_count = 0` (a filter, not a verdict — see `acted_count`), because a run with uncountable effects has an INCOMPLETE acted count, not a zero one. Null on rows written before this was tracked.',
       group: 'Outcome',
     }),