Merge pull request #134 from zestones/133-m95-full-frontend-polish

feat: update demo build spec for improved clarity and accuracy in onb…

Author: zestones

backend/main.py

@@ -137,9 +137,7 @@ async def health() -> dict[str, str]:
     # to ``/mcp/<secret>``.
     from modules.sandbox.app import sandbox_app
 
-    app.mount(
-        f"/sandbox/{settings.aria_mcp_path_secret}", sandbox_app, name="sandbox-secret"
-    )
+    app.mount(f"/sandbox/{settings.aria_mcp_path_secret}", sandbox_app, name="sandbox-secret")
 
     return app
 

backend/modules/debug/router.py

@@ -51,9 +51,7 @@ async def replay_investigator(
         work_order_id,
     )
     if wo is None:
-        raise HTTPException(
-            status_code=404, detail=f"work_order {work_order_id} not found"
-        )
+        raise HTTPException(status_code=404, detail=f"work_order {work_order_id} not found")
 
     await conn.execute(
         "UPDATE work_order SET status = 'detected' WHERE id = $1",

backend/tests/unit/agents/investigator/test_prompts.py

@@ -63,8 +63,7 @@ def test_sandbox_section_contains_at_least_two_worked_examples() -> None:
     """
     fence_count = SANDBOX_DIAGNOSTICS_SECTION.count("PY")
     assert fence_count >= 4, (
-        f"expected at least 2 Python heredoc examples (>=4 'PY' fences), "
-        f"found {fence_count}"
+        f"expected at least 2 Python heredoc examples (>=4 'PY' fences), " f"found {fence_count}"
     )
 
 
@@ -76,9 +75,9 @@ def test_sandbox_section_references_fft_and_trend_techniques() -> None:
     lower = SANDBOX_DIAGNOSTICS_SECTION.lower()
     assert "fft" in lower, "FFT example is missing"
     # polyfit / linear regression wording — either form is acceptable.
-    assert any(term in lower for term in ("polyfit", "regression", "slope")), (
-        "linear-trend / regression example is missing"
-    )
+    assert any(
+        term in lower for term in ("polyfit", "regression", "slope")
+    ), "linear-trend / regression example is missing"
 
 
 # ---------------------------------------------------------------------------
@@ -91,9 +90,9 @@ def test_sandbox_section_mandates_sandbox_colon_prefix() -> None:
     so the numerical evidence lands in the work-order text itself.
     Dropping this rule makes the capability indistinguishable from
     Messages API token arithmetic in the RCA surface."""
-    assert "Sandbox:" in SANDBOX_DIAGNOSTICS_SECTION, (
-        "the Sandbox: prefix rule for submit_rca.root_cause is missing"
-    )
+    assert (
+        "Sandbox:" in SANDBOX_DIAGNOSTICS_SECTION
+    ), "the Sandbox: prefix rule for submit_rca.root_cause is missing"
 
 
 def test_sandbox_section_has_failure_mode_keyed_rules() -> None:
@@ -109,28 +108,28 @@ def test_sandbox_section_has_failure_mode_keyed_rules() -> None:
 def test_sandbox_section_mandates_render_sandbox_execution_call() -> None:
     """The visible-proof card must be called after bash + before submit_rca.
     Without this line the artifact is orphaned."""
-    assert "render_sandbox_execution" in SANDBOX_DIAGNOSTICS_SECTION, (
-        "render_sandbox_execution mandate is missing from the prompt"
-    )
+    assert (
+        "render_sandbox_execution" in SANDBOX_DIAGNOSTICS_SECTION
+    ), "render_sandbox_execution mandate is missing from the prompt"
 
 
 def test_sandbox_section_references_required_regression_metrics() -> None:
     """Drift-class rule requires slope / r_squared / eta — all three must
     appear so the agent knows the expected Sandbox: line shape."""
     for metric in ("slope_per_hour", "r_squared", "eta_"):
-        assert metric in SANDBOX_DIAGNOSTICS_SECTION, (
-            f"drift-class metric {metric!r} is missing from the prompt"
-        )
+        assert (
+            metric in SANDBOX_DIAGNOSTICS_SECTION
+        ), f"drift-class metric {metric!r} is missing from the prompt"
 
 
 def test_sandbox_section_references_correlation_metrics() -> None:
     """Coupling-class rule requires rho / n so the agent's output is
     parseable in the Sandbox: prefix."""
     lower = SANDBOX_DIAGNOSTICS_SECTION.lower()
     assert "rho" in lower, "coupling-class rho metric is missing"
-    assert any(token in lower for token in ("n_samples", "n=")), (
-        "coupling-class sample-count metric is missing"
-    )
+    assert any(
+        token in lower for token in ("n_samples", "n=")
+    ), "coupling-class sample-count metric is missing"
 
 
 # ---------------------------------------------------------------------------

backend/tests/unit/agents/test_sentinel.py

@@ -736,10 +736,7 @@ async def test_forecast_tick_skips_flat_series(patch_forecast) -> None:
     from datetime import timedelta
 
     now = datetime.now(timezone.utc)
-    flat = [
-        {"time": now - timedelta(minutes=i), "raw_value": 10.0}
-        for i in range(60, 0, -1)
-    ]
+    flat = [{"time": now - timedelta(minutes=i), "raw_value": 10.0} for i in range(60, 0, -1)]
     signals = [
         {
             "signal_def_id": 10,

backend/tests/unit/agents/test_ui_tools.py

@@ -42,9 +42,9 @@ def test_all_llm_render_tools_are_anthropic_format():
     """Every tool in ALL_LLM_RENDER_TOOLS is a valid Anthropic tool descriptor."""
     # 7 original LLM render tools (render_correlation_matrix dropped per audit
     # §2) + render_sandbox_execution (M5.7 / #105) = 8.
-    assert len(ALL_LLM_RENDER_TOOLS) == 8, (
-        "Expected 8 LLM render tools: 7 original + render_sandbox_execution (M5.7 / #105)"
-    )
+    assert (
+        len(ALL_LLM_RENDER_TOOLS) == 8
+    ), "Expected 8 LLM render tools: 7 original + render_sandbox_execution (M5.7 / #105)"
     for tool in ALL_LLM_RENDER_TOOLS:
         _assert_anthropic_format(tool)
 

backend/tests/unit/modules/sandbox/test_csv.py

@@ -89,9 +89,7 @@ async def _collect_body(response: Any) -> bytes:
 @pytest.mark.asyncio
 async def test_happy_path_returns_csv_ordered_ascending() -> None:
     t0 = datetime(2026, 4, 24, 12, 0, 0, tzinfo=timezone.utc)
-    rows = [
-        _sample_row(t0 + timedelta(seconds=30 * i), 2.2 + 0.01 * i) for i in range(5)
-    ]
+    rows = [_sample_row(t0 + timedelta(seconds=30 * i), 2.2 + 0.01 * i) for i in range(5)]
     conn = _FakeConn(definition=_sample_definition(), rows=rows)
 
     response = await sandbox_mod.signal_csv(