features/steps/materializers_coverage_boost_steps.py

@@ -279,6 +279,7 @@ def step_given_mock_snapshot(context: Context) -> None:
     context.mock_snapshot = StructuredOutput(
         command="test",
         session_id="ses-000001",
+        status="ok",
         elements=[],
         exit_code=0,
         metadata={},

features/steps/output_rendering_steps.py

@@ -552,7 +552,7 @@ def step_json_valid(context: Context) -> None:
 @then('the json output contains element type "{etype}"')
 def step_json_element_type(context: Context, etype: str) -> None:
     data = context.json_data
-    elements = data.get("elements", [])
+    elements = data.get("data", [])
     types = [e.get("type") for e in elements]
     assert etype in types, f"Expected element type {etype!r} in {types}"
 

robot/helper_output_rendering.py

@@ -60,8 +60,10 @@ def test_json_output() -> None:
         table.close()
     output = strategy.get_output()
     data = json.loads(output)
-    assert "elements" in data
-    assert len(data["elements"]) == 2
+    assert "data" in data
+    assert len(data["data"]) == 2
+    assert "status" in data
+    assert data["status"] == "ok"
     print("output-rendering-json-output-ok")
 
 
@@ -74,7 +76,9 @@ def test_yaml_output() -> None:
         panel.close()
     output = strategy.get_output()
     data = yaml.safe_load(output)
-    assert "elements" in data
+    assert "data" in data
+    assert "status" in data
+    assert data["status"] == "ok"
     print("output-rendering-yaml-output-ok")
 
 
@@ -271,7 +275,7 @@ def test_json_all_elements() -> None:
         session.action_hint(["cmd"])
     output = strategy.get_output()
     data = json.loads(output)
-    types = [e["type"] for e in data["elements"]]
+    types = [e["type"] for e in data["data"]]
     expected = [
         "panel",
         "table",
@@ -327,7 +331,7 @@ def test_yaml_all_elements() -> None:
         session.action_hint(["cmd"])
     output = strategy.get_output()
     data = yaml.safe_load(output)
-    types = [e["type"] for e in data["elements"]]
+    types = [e["type"] for e in data["data"]]
     expected = [
         "panel",
         "table",

src/cleveragents/cli/output/__init__.py

@@ -65,11 +65,9 @@ yet needed in the current milestone.
   The spec does not define an element count limit.  A soft limit of
   10,000 is enforced as a DoS guard.
 
-**SD-13 (JSON/YAML envelope structure)**
-  The spec defines ``status``/``data`` wrapper, ``messages``, and
-  ``timing`` fields in the serialised envelope.  This implementation
-  serialises a flat ``StructuredOutput`` dict with ``command``,
-  ``session_id``, ``exit_code``, ``elements``, and ``metadata``.
+**SD-13 (JSON/YAML envelope structure — IMPLEMENTED)**
+  The spec's envelope fields (``status``, ``data``, ``messages``,
+  ``timing``) are now emitted.  Implemented per M4 fix (#884).
 
 **SD-14 (NO_COLOR env var — IMPLEMENTED)**
   ``NO_COLOR`` (https://no-color.org/) is now respected: when set,

src/cleveragents/cli/output/materializers.py

@@ -275,16 +275,18 @@ def _snapshot_to_dict(snapshot: StructuredOutput) -> dict[str, Any]:
 
     P1-2 fix (Luis review): includes the ``timing`` field when present,
     matching the spec's JSON envelope structure (§27022).
+
+    M4 fix: renamed ``elements`` → ``data``, ``metadata`` → ``messages``,
+    and added ``status`` field to match the specification envelope.
     """
     result: dict[str, Any] = {
         "command": snapshot.command,
-        "session_id": snapshot.session_id,
+        "status": snapshot.status,
         "exit_code": snapshot.exit_code,
-        "elements": [_element_to_dict(e) for e in snapshot.elements],
-        "metadata": snapshot.metadata,
+        "data": [_element_to_dict(e) for e in snapshot.elements],
+        "timing": snapshot.timing,
+        "messages": snapshot.metadata,
     }
-    if snapshot.timing is not None:
-        result["timing"] = snapshot.timing
     return result
 
 

src/cleveragents/cli/output/session.py

@@ -72,6 +72,7 @@ class StructuredOutput(BaseModel):
 
     command: str
     session_id: str
+    status: str = "ok"
     elements: list[ElementSnapshot] = Field(default_factory=list)
     exit_code: int = 0
     timing: dict[str, Any] | None = None
@@ -424,12 +425,17 @@ class OutputSession:
 
         P1-2 fix (Luis review): includes ``timing`` when available so
         that accumulate strategies (JSON/YAML) emit the timing field.
+
+        M4 fix: derives ``status`` from ``exit_code`` — ``"ok"`` when
+        zero, ``"error"`` otherwise.
         """
         with self._lock:
             elements = [h.element_copy() for h in self._handles.values()]
+        status = "error" if self._exit_code != 0 else "ok"
         snap = StructuredOutput(
             command=self.command,
             session_id=self.session_id,
+            status=status,
             elements=list(elements),
             exit_code=self._exit_code,
             metadata=dict(self._metadata),
@@ -445,9 +451,11 @@ class OutputSession:
                 # Already closed — return snapshot with cached timing
                 # (R2-C2 fix: previously timing was None on double-close).
                 elements = [h.element_copy() for h in self._handles.values()]
+                status = "error" if self._exit_code != 0 else "ok"
                 snap = StructuredOutput(
                     command=self.command,
                     session_id=self.session_id,
+                    status=status,
                     elements=list(elements),
                     exit_code=self._exit_code,
                     metadata=dict(self._metadata),