feat(tasks): add skill-templated task graph execution

app-instance/backend/tests/unit/test_skill_learning_eval.py

@@ -395,6 +395,52 @@ def test_replay_main_score_uses_validator_not_tool_success(tmp_path: Path) -> No
     assert report.synthetic_score_avg is not None
 
 
+def test_replay_real_case_without_validator_uses_same_output_scoring_for_both_arms(tmp_path: Path) -> None:
+    pipeline = _pipeline(tmp_path, task_score=0.8)
+    pipeline.learning_store.update_learning_candidate(
+        "candidate-1",
+        evidence={
+            "eval_cases": [
+                {
+                    "run_id": "real-no-validator",
+                    "task_id": "real-no-validator",
+                    "session_id": "eval",
+                    "task_text": "Summarize the release checklist.",
+                    "accepted_score": 0.8,
+                }
+            ]
+        },
+    )
+    draft = pipeline.draft_service.create_new_skill_draft(
+        skill_name="release-checklist",
+        proposed_content="# Release\n\nRun tests.",
+        proposed_frontmatter={"description": "release", "tools": []},
+        created_by="test",
+        reason="test",
+    )
+    pipeline.learning_store.update_learning_candidate("candidate-1", draft_skill_name=draft.skill_name, draft_id=draft.draft_id)
+
+    report = asyncio.run(
+        pipeline.evaluate_draft(
+            "candidate-1",
+            draft.skill_name,
+            draft.draft_id,
+            provider_bundle=_bundle(),
+            replay_runner=FakeReplayRunner(
+                baseline_answer="Release checklist summarized.",
+                candidate_answer="Release checklist summarized.",
+            ),
+        )
+    )
+
+    case = next(item for item in report.case_reports if item["run_id"] == "real-no-validator")
+    legacy_case = next(item for item in report.cases if item["run_id"] == "real-no-validator")
+    assert case["baseline_score"] == 0.7
+    assert case["candidate_score"] == 0.7
+    assert case["delta"] == 0.0
+    assert legacy_case["delta"] == 0.0
+
+
 def test_synthetic_cases_without_validator_are_not_replay_scored(tmp_path: Path) -> None:
     pipeline = _pipeline(tmp_path)
     pipeline.learning_store.update_learning_candidate(